{ "row": "mit-ai-risk-repository", "entity": "mit-ai-risk-repository", "address": "mit ai risk repository", "addresses": [ { "address": "mit ai risk repository", "slug": "mit-ai-risk-repository-aio-20261029", "addr_id": "ADDR-5762c6084031", "obs_id": "OBS-21a7cea64b70", "axn": "AXN:0AAF.CAPTURE.🚀🕌🔒🏠✖️🌉" } ], "epoch": "2026-10-29", "type": "C (public entity on which the archive bears: a database; an entity of the MMRS meta-aggregator)", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "status": { "procedure": "EA-NEGONT-02 v0.9, #1690: equal case, ρ = extends, read from the MMRS record and the capture record; archive selected by D/R/O (zero candidates in 1,711 texts; no hop; O-class matched nothing), with the MMRS record (26 claims) and the Capture Registry's findings and analysis (6 claims) ledgered as the archive's records of the entity; the field B by three origins (10 card sources for 10 cards, 2 own pages not surfaced, Wikipedia: no article); the composer read the ledgers, never the transcript, except where the brief asks for T's claims and the delta", "ledger": "unaudited (§3.8: one extractor per ledger); field quotes 59/59 verified by script against saved text or card text; archive quotes 32/32 verified verbatim against the registry, census.json and articles.py; carried as marked: MMRS-01, MMRS-02, MMRS-13, MMRS-20, CAP-1-02, CAP-1-04 contested", "frozen": false, "breaches": [] }, "spec": "EA-NEGONT-02 v0.9, #1690", "objects": { "T": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/T1-mit-ai-risk-repository-aio-20261029.txt", "sha256": "15d8ca3b8664862cbe68d24e00567a25a7f07e97fd747c753a1ca01e53a669d0", "words": 207, "cards": 10, "words_note": "one transcript at one address (207 words: the composed body to the closer, inline link URLs and citation brackets removed, card rail excluded; counted with [A-Za-z0-9'’-]+); 10 cards", "claims": [ "T1-1 definition: 'a living, open-access database developed by MIT FutureTech that catalogs and standardizes artificial intelligence risks from dozens of global frameworks' (markers 1–3: the entity link's token and two others)", "T1-2 scale: 'The database compiles over 1,700 distinct AI risks extracted from more than 65 published frameworks, industry guidelines, and policy documents.'", "T1-3 causal taxonomy: 'Classifies risks based on the responsible entity (human vs. AI), intent (intentional vs. unintentional), and timing (pre-deployment vs. post-deployment).'", "T1-4 finding: 'Findings indicate that roughly 38% to 42% of risks stem directly from AI systems, while a large share originates from human decisions during design and deployment.'", "T1-5 domain taxonomy: 'Organizes risks into 7 core domains and 24 subdomains', the seven listed (Discrimination & toxicity … AI system safety, failures, & limitations)", "T1-6 connected tool: 'MIT AI Risk Navigator: An interactive web tool that links the repository to real-world AI incidents, global governance frameworks (such as the EU AI Act), and risk mitigations.'", "T1-7 connected tool: 'Incident Tracker: Collaborates with the AI Incident Database to connect abstract risks to documented real-world harms.'", "T1-8 closer: 'Would you like to explore how to use the repository for a specific industry audit, regulatory compliance check (such as the EU AI Act), or risk assessment?'" ], "register": { "register": "data/EA-WG-CAPTURES-01.json", "slug": "mit-ai-risk-repository-aio-20261029", "obs_id": "OBS-21a7cea64b70" }, "transcripts": [ { "slug": "mit-ai-risk-repository-aio-20261029", "address": "mit ai risk repository", "addr_id": "ADDR-5762c6084031", "obs_id": "OBS-21a7cea64b70", "axn": "AXN:0AAF.CAPTURE.🚀🕌🔒🏠✖️🌉", "date": "2026-10-29", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "sha256": "15d8ca3b8664862cbe68d24e00567a25a7f07e97fd747c753a1ca01e53a669d0", "sha256_of": "the registry's 'transcript' string (utf-8)", "sha256_raw": "eb27b755b712a52fa7e3ed5675a7e4f53383a72040ea86aa35f46649feef5679", "words": 207, "cards": 10, "card_rail": [ "1. MIT AI Risk Initiative — MIT AI Risk Initiative", "2. MIT AI Risk Initiative — AI Risk Repository Report updated (April 2025)", "3. MIT Initiative on the Digital Economy — MIT Researchers Create an AI Risk Repository - MIT Initiative on the Digital Economy", "4. MIT FutureTech — FutureTech", "5. MIT AI Risk Navigator — MIT AI Risk Navigator", "6. arXiv — [2408.12622] The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence", "7. MIT AI Risk Initiative — Repository Update: December 2025", "8. MIT AI Risk Initiative — About", "9. LinkedIn — MIT AI Risk Initiative | LinkedIn", "10. ScienceDirect.com — The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence - ScienceDirect" ], "claims": [ "T1-1 definition: 'a living, open-access database developed by MIT FutureTech that catalogs and standardizes artificial intelligence risks from dozens of global frameworks' (markers 1–3: the entity link's token and two others)", "T1-2 scale: 'The database compiles over 1,700 distinct AI risks extracted from more than 65 published frameworks, industry guidelines, and policy documents.'", "T1-3 causal taxonomy: 'Classifies risks based on the responsible entity (human vs. AI), intent (intentional vs. unintentional), and timing (pre-deployment vs. post-deployment).'", "T1-4 finding: 'Findings indicate that roughly 38% to 42% of risks stem directly from AI systems, while a large share originates from human decisions during design and deployment.'", "T1-5 domain taxonomy: 'Organizes risks into 7 core domains and 24 subdomains', the seven listed (Discrimination & toxicity … AI system safety, failures, & limitations)", "T1-6 connected tool: 'MIT AI Risk Navigator: An interactive web tool that links the repository to real-world AI incidents, global governance frameworks (such as the EU AI Act), and risk mitigations.'", "T1-7 connected tool: 'Incident Tracker: Collaborates with the AI Incident Database to connect abstract risks to documented real-world harms.'", "T1-8 closer: 'Would you like to explore how to use the repository for a specific industry audit, regulatory compliance check (such as the EU AI Act), or risk assessment?'" ], "markers": "inline links are opaque google.com/goto tokens; the entity's name is itself linked (the same token as marker 1 of the definition); the Navigator is linked by name; the cards are not mapped to markers by the paste", "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/T1-mit-ai-risk-repository-aio-20261029.txt" } ] }, "KO": { "title": "The MIT AI Risk Repository", "genre": "knowledge object, equal case, under EA-NEGONT-02 v0.9 (#1690): the archive's sources and the field's admitted on equal terms, ontology included, in one claim space", "status": "composed 2026-10-29 under v0.9 (brief rebuild/negative-of-the-negative/v09/BRIEF-NEW-ROW-AGGREGATOR.md); not frozen; ledgers unaudited", "sentences": [ { "n": 1, "para": 0, "sense": "what it is", "text": "The MIT AI Risk Repository is a database of more than 1,700 AI risks extracted from published risk frameworks, run by the MIT AI Risk Initiative (MIT FutureTech) (MMRS record); its own page gives it three parts, an AI Risk Database that \"captures 1700+ risks extracted from 74 existing frameworks and classifications of AI risks\", a Causal Taxonomy and a Domain Taxonomy, and the Initiative's home page calls it a \"living repository of AI risks\".", "claims": [ "MMRS-01", "F44", "F2" ] }, { "n": 2, "para": 0, "sense": "the unit", "text": "Its unit is the risk as a framework states it (MMRS record), \"risk, extracted from published risk frameworks\" in the record's field, each risk linked to \"the source information (paper title, authors) and supporting evidence, such as quotes and page numbers\" (MIT IDE), and each source framework summarized with its source documents and key excerpts published (MMRS record), as the home page's deck of the 65 Version 3 frameworks gives them.", "claims": [ "MMRS-03", "MMRS-12", "F12", "MMRS-04", "F4" ] }, { "n": 3, "para": 0, "sense": "what it is", "text": "Its report post of April 2025 calls it \"a structured and evolving resource\" consolidating how AI risks are categorized across academic, industry and policy literatures, its LinkedIn page \"a comprehensive living database of over 1000 AI risks categorized by their cause and risk domain\", and the Initiative's file for machines names /risks its canonical page, with \"methodology notes, use cases, limitations, data access links, and citation guidance\".", "claims": [ "F9", "F42", "F59" ] }, { "n": 4, "para": 1, "sense": "purpose", "text": "Slattery and colleagues start from terminology: \"researchers, policymakers, and technology companies lack shared terminology for discussing AI risks\", so that \"privacy\" names a model's ability to leak sensitive training data in one framework and freedom from government surveillance in another (arXiv:2408.12622, v3).", "claims": [ "F25", "F26" ] }, { "n": 5, "para": 1, "sense": "purpose", "text": "The Initiative offers \"Authoritative data and frameworks to help you identify, prioritize, and manage the risks from AI\" (home page); the v1 abstract calls the repository, to its authors' knowledge, \"the first attempt to rigorously curate, analyze, and extract AI risk frameworks into a publicly accessible, comprehensive, extensible, and categorized risk database\", and the published article's snippet promises \"concrete tools for managing AI's societal impact across multiple stakeholder[s]\" (ScienceDirect card).", "claims": [ "F1", "F21", "F43" ] }, { "n": 6, "para": 1, "sense": "what it gathers", "text": "The MMRS record reads the same object as an aggregator of aggregators, \"its sources are frameworks, which are themselves syntheses\": \"What enters is what a framework named\", and \"A risk no framework has written down has no row\"; the repository's FAQ states the boundary as its limit, 65 documents from more than 17,000 records screened, which \"May be missing emerging, domain-specific risks, and unpublished risks\", and a framework can itself reach back to events, as the taxonomy of generative AI failures built from 499 publicly reported incidents that Version 4 added (December 2025 update).", "claims": [ "MMRS-06", "MMRS-07", "MMRS-08", "F49", "F33" ] }, { "n": 7, "para": 2, "sense": "method", "text": "Its method is \"a systematic review of taxonomies and other structured classifications of AI risk followed by an expert consultation\", its taxonomies made by \"a best-fit framework synthesis\" (v1 abstract), restated in the FAQ as a systematic search with forwards and backwards searching and expert consultation that identified 65 documents and extracted 1600+ risks; there are two taxonomies because causal categories such as \"Diffusion of responsibility\" do not map to hazard categories such as \"Misuse\" (FAQ).", "claims": [ "F18", "F47", "F53" ] }, { "n": 8, "para": 3, "sense": "the Causal Taxonomy", "text": "The Causal Taxonomy classifies each risk by Entity (Human, AI), Intentionality (Intentional, Unintentional) and Timing (Pre-deployment, Post-deployment) (v1 abstract), and the repository's page gives each factor a third value, \"Other: Due to some other reason or ambiguous\" for entity.", "claims": [ "F19", "F46" ] }, { "n": 9, "para": 3, "sense": "the Domain Taxonomy", "text": "The Domain Taxonomy sorts risks into seven domains, discrimination and toxicity, privacy and security, misinformation, malicious actors and misuse, human-computer interaction, socioeconomic and environmental, and AI system safety, failures and limitations (v1 abstract), each with subdomains (MMRS record): 23 in v1, 24 on the repository's page, 25 in the April 2025 report post, the subdomain for multi-agent risks added in April 2025 (MMRS record; report post).", "claims": [ "F20", "MMRS-05", "F45", "F8", "F7" ] }, { "n": 10, "para": 4, "sense": "findings", "text": "Its v3 finding is that, \"contrary to common assumptions, human decisions cause nearly as many AI risks (38%) as the AI systems themselves (42%)\" (arXiv v3); the first release had found 51% of risks caused by AI systems and 34% by humans, 65% after deployment and 10% before, and Slattery gave as its key finding \"the average framework covering only 34% of the identified risk subdomains and even the most significant frameworks covering only 70%\" (MIT IDE).", "claims": [ "F28", "F14", "F13" ] }, { "n": 11, "para": 5, "sense": "count and version", "text": "The counts are versioned: 777 risks from 43 frameworks at the first release, \"A new MIT FutureTech project\" (MIT IDE, 21 August 2024), \"777 risks extracted from 43 taxonomies\" in the v1 abstract; 22 frameworks added to 65 documents and 1,612 unique risk entries in April 2025 (report post); Version 4 live on 4 December 2025 with 9 frameworks and about 200 risk categories added, \"over 1,700 coded risks\" (December 2025 update); and \"74 frameworks containing 1,725 distinct risks\" in v3 of the paper, repeated on the repository's page as 1700+ from 74.", "claims": [ "F11", "F17", "F6", "F31", "F32", "F27", "F44" ] }, { "n": 12, "para": 5, "sense": "count and version", "text": "Older figures stand beside the current one on the Initiative's own pages: \"1,700+ AI risks from 65 frameworks\" on the home page, \"more than 1,600 risks from more than 65 frameworks and papers\" and, further down the same About page, \"1700+ risks from 70+ frameworks\", and \"more than 1,700\" in the LinkedIn post; the MMRS record gives \"more than 1,700 risks from 65 frameworks\" as of 2026-10-28, read from airisk.mit.edu for \"scope, version\", with lifecycle \"version 4, December 2025\" (MMRS record).", "claims": [ "F3", "F37", "F38", "F41", "MMRS-13", "MMRS-16", "MMRS-15" ] }, { "n": 13, "para": 5, "sense": "count and version", "text": "The Capture Registry takes 1,700 from 65 as the record's count for version 4 and finds that, of the three sizes on the cards, \"the composition gives the current one\" (Capture Registry, 2026-10-29); version 4's post adds 9 frameworks to version 3's 65, and v3 of the paper and the repository's page give 74 while the home page gives 65, so the record's 65 is held beside the 74, each with its source.", "claims": [ "CAP-1-02", "CAP-1-04", "F32", "F6", "F27", "F44", "F3" ] }, { "n": 14, "para": 6, "sense": "stated limits", "text": "Its stated limits are \"a single expert reviewer for extraction and coding\", risks extracted as presented, poorly communicated ones included, and taxonomies that \"prioritize clarity and simplicity over nuance\" and do not categorize risks by impact, likelihood or their interaction (repository FAQ); Slattery expected it to \"need some modification before it is useful for most audiences\" (MIT IDE).", "claims": [ "F48", "F50", "F15" ] }, { "n": 15, "para": 6, "sense": "the record kept apart", "text": "The MMRS record lists extraction method and licence as unchecked; the field states both, the method above and \"Data from the MIT AI Risk Initiative is licensed under CC BY 4.0\" (home page), \"most of our outputs are published under a CC BY 4.0 licence\" (About page), \"Our Database is free to copy and use\" (repository page), the paper CC BY 4.0 in Patterns (arXiv).", "claims": [ "MMRS-17", "F5", "F40", "F51", "F30" ] }, { "n": 16, "para": 7, "sense": "operator", "text": "Its operator is the MIT AI Risk Initiative (AIRI), \"based at MIT FutureTech, a joint initiative of the Institute of the Digital Economy and MIT CSAIL\" (About page), as the MMRS record gives it, on the operator's statement.", "claims": [ "F34", "F35", "MMRS-11" ] }, { "n": 17, "para": 7, "sense": "the paper", "text": "Its paper is by Peter Slattery, Alexander K. Saeri, Emily A. C. Grundy, Jess Graham, Michael Noetel, Risto Uuk, James Dao, Soroush Pour, Stephen Casper and Neil Thompson, submitted on 14 August 2024 and revised to v3 on 5 May 2026 (arXiv), with versions of August 2024 and April 2025 (report post), published in Patterns (Cell Press, 2026), and the repository asks to be cited by its preprint (FAQ).", "claims": [ "F29", "F10", "F30", "F54" ] }, { "n": 18, "para": 8, "sense": "the Initiative's workstreams", "text": "The repository is one of the Initiative's \"several connected workstreams, each answering a different question\", beside an AI Incident Tracker \"Developed in collaboration with the AI Incident Database\" (About page), which the Initiative's LinkedIn post calls \"a collaboration with the Responsible AI Collaborative, which connects risks to over 1,400 incidents\".", "claims": [ "F36", "F41" ] }, { "n": 19, "para": 8, "sense": "the Initiative's workstreams", "text": "The MMRS record keeps the incidents apart from the risks, \"real-world incidents are kept separately in its AI Incident Tracker\", records the tracker as part of the repository, classifying AI Incident Database reports against the same taxonomy through Arcola AI's language-model pipeline, and reads that layer as \"a second-order reading of the AI Incident Database, done by machine\" (MMRS record); the field's About page sets the tracker beside the repository as a workstream, and the two placements stand each with its source.", "claims": [ "MMRS-14", "MMRS-20", "MMRS-02", "MMRS-21", "MMRS-09", "F36" ] }, { "n": 20, "para": 8, "sense": "place", "text": "In the record's harm basin the repository \"sits over the harm basin as a taxonomy for it\", its role harm and its neighbour the AI Incident Database, whose own record says the repository \"aggregates risk frameworks over the same ground and keeps its own incident tracker\" (MMRS record); the repository's FAQ names as other databases of AI risks MITRE ATT&CK, airisk.io, AVID and aitracker.org.", "claims": [ "MMRS-10", "MMRS-18", "MMRS-19", "MMRS-22", "F52" ] }, { "n": 21, "para": 9, "sense": "the Navigator", "text": "The MIT AI Risk Navigator, \"A project of the AI Risk Initiative at MIT FutureTech\" with the support of the Cambridge Boston Alignment Initiative, connects the repository's datasets \"through a shared taxonomy\" and says of itself \"This is a research prototype. The data and analyses are preliminary and not yet validated\" (its home page); the Capture Registry finds it \"a third MIT surface\" while \"the census holds the repository and the tracker\", \"Added as a candidate\", and the MMRS record lists it as a candidate to enter \"once it has been read at its own pages\", a distinct entity.", "claims": [ "F24", "F22", "F23", "CAP-1-05", "CAP-1-06", "MMRS-23", "MMRS-24" ] }, { "n": 22, "para": 10, "sense": "machine-facing guidance", "text": "The Initiative publishes a file \"to orient LLMs and AI agents\" that asks them to \"Refer to the project as the MIT AI Risk Initiative\", to \"Check the linked pages before giving exact counts, dates, version numbers, or coverage claims\", to \"distinguish reported incidents from predicted risks\", to note that the Navigator \"is a research prototype\", and not to present the datasets \"as legal, safety, or investment advice\" (llms.txt).", "claims": [ "F55", "F56", "F58", "F57" ] }, { "n": 23, "para": 10, "sense": "reach", "text": "Its About page counts 250+ citations, the 2025 and 2026 International AI Safety Reports among them, and 290,000+ website users, and calls the site \"One of the top sources returned for search queries on AI risks and incidents\"; in its first week over 35,000 people used the website (MIT IDE).", "claims": [ "F39", "F16" ] }, { "n": 24, "para": 11, "sense": "the record kept apart", "text": "The MMRS record has \"two nodes, the MMRS record and the entity it describes\", each assertion with a source and a basis and the lifecycle \"kept apart from the observation that saw it\", and it links the capture as \"the composition layer's account of the entity, seated in the Capture Registry\" (MMRS record).", "claims": [ "MMRS-25", "MMRS-26" ] }, { "n": 25, "para": 11, "sense": "reception", "text": "At the entity's name the Capture Registry finds \"The repository is composed at a finer grain than the record holds, its taxonomy named domain by domain\", the seven domains and the causal taxonomy \"the record's 'domain taxonomy with subdomains' given at full grain\", checked against the record and the entity's own pages on 2026-10-29 (Capture Registry, 2026-10-29).", "claims": [ "CAP-1-01", "CAP-1-02", "CAP-1-03" ] } ], "joins": [ { "act": "order of senses", "sentences": "1–25", "claims": [], "note": "equal case (§4.7): the entity's dependency order, field and archive claims ordered together by genealogy within each sense; neither ledger placed first as a block. What it is and its unit (¶0) precede purpose and what the selection gathers (¶1), since what it gathers depends on the unit; the method (¶2) precedes the taxonomies it produced (¶3) and their findings (¶4); the counts by version (¶5) follow, then the limits and the record's unchecked items (¶6), the operator and the paper (¶7), the workstreams, the tracker and the place in the harm basin (¶8), the Navigator (¶9), the entity's own statements to machines and its reach (¶10), and the record and the composition layer's observation (¶11), as relations of the entity already composed." }, { "act": "holding apart (contested count)", "sentences": [ 12, 13 ], "claims": [ "MMRS-01", "MMRS-13", "CAP-1-02", "CAP-1-04", "F3", "F27", "F32", "F44" ], "note": "P_coh: the record's and the registry's 65 and the field's 74 composed each with its source; the composer states the arithmetic the field gives (9 added to 65) and does not rule which page is current (open block)" }, { "act": "holding apart (the tracker's place)", "sentences": [ 19 ], "claims": [ "MMRS-20", "MMRS-02", "MMRS-14", "F36", "F41" ], "note": "P_coh: the record's part_of and the field's sibling workstream composed each with its source; no coordinating claim" }, { "act": "holding together (what it gathers and its limit)", "sentences": [ 6 ], "claims": [ "MMRS-06", "MMRS-07", "MMRS-08", "F49", "F33" ], "note": "the record's reading of the unit and the FAQ's stated limit composed side by side, the reading at interpretation, the limit at self-description; the framework built from incidents placed beside them as the field's own counter-instance" }, { "act": "holding together (subdomain counts)", "sentences": [ 9 ], "claims": [ "F20", "F45", "F8" ], "note": "three counts on MIT's own pages composed each with its source (open block)" } ], "plan": { "order": [ "MMRS-01", "F44", "F2", "MMRS-03", "MMRS-12", "F12", "MMRS-04", "F4", "F9", "F42", "F59", "F25", "F26", "F1", "F21", "F43", "MMRS-06", "MMRS-07", "MMRS-08", "F49", "F33", "F18", "F47", "F53", "F19", "F46", "F20", "MMRS-05", "F45", "F8", "F7", "F28", "F14", "F13", "F11", "F17", "F6", "F31", "F32", "F27", "F3", "F37", "F38", "F41", "MMRS-13", "MMRS-16", "MMRS-15", "CAP-1-02", "CAP-1-04", "F48", "F50", "F15", "MMRS-17", "F5", "F40", "F51", "F30", "F34", "F35", "MMRS-11", "F29", "F10", "F54", "F36", "MMRS-14", "MMRS-20", "MMRS-02", "MMRS-21", "MMRS-09", "MMRS-10", "MMRS-18", "MMRS-19", "MMRS-22", "F52", "F24", "F22", "F23", "CAP-1-05", "CAP-1-06", "MMRS-23", "MMRS-24", "F55", "F56", "F58", "F57", "F39", "F16", "MMRS-25", "MMRS-26", "CAP-1-01", "CAP-1-03" ], "senses": { "what it is": [ 1, 3 ], "the unit": [ 2 ], "purpose": [ 4, 5 ], "what it gathers": [ 6 ], "method": [ 7 ], "the Causal Taxonomy": [ 8 ], "the Domain Taxonomy": [ 9 ], "findings": [ 10 ], "count and version": [ 11, 12, 13 ], "stated limits": [ 14 ], "the record kept apart": [ 15, 24 ], "operator": [ 16 ], "the paper": [ 17 ], "the Initiative's workstreams": [ 18, 19 ], "place": [ 20 ], "the Navigator": [ 21 ], "machine-facing guidance": [ 22 ], "reach": [ 23 ], "reception": [ 25 ] }, "case_by_sense": { "equal (one claim space; nothing translated)": [ "what it is", "the unit", "purpose", "what it gathers", "method", "the Causal Taxonomy", "the Domain Taxonomy", "findings", "count and version", "stated limits", "the record kept apart", "operator", "the paper", "the Initiative's workstreams", "place", "the Navigator", "machine-facing guidance", "reach", "reception" ] }, "carried": { "archive": 32, "field": 59, "not_composed": {} }, "conflict_log": [ { "rule": "D_pres vs channel", "result": "no cut: the KO carries every field and archive claim; the compression's bound met (body 299 words) with no channel log" }, { "rule": "P_coh", "result": "MMRS-01/13 and CAP-1-02/04 against F27, F44 (sentences 12–13); MMRS-20, MMRS-02 against F36, F41 (sentence 19); F20, F45, F8 (sentence 9): each composed with its source" }, { "rule": "N_c", "result": "no claim lowered by standing; the MMRS record's readings (MMRS-06–10, MMRS-14, MMRS-22) carried at interpretation, as the record states them; the Capture Registry's findings at their own modality" }, { "rule": "C_ex", "result": "senses composed ⊇ senses in the ledgers: 19 senses, every ledger lineage present" } ] } }, "P": { "title": "MIT AI Risk Repository", "genre": "the compression: AIO's interaction grammar (lede, headed clusters, bolded labels, card rail), composed from the v0.9 plan in one claim space; it faces on the card, with AIO's transcript a record within it (§4.8, §5.8)", "status": "composed 2026-10-29 from the v0.9 knowledge object; not frozen", "lede": { "text": "is a living database of more than 1,700 AI risks extracted from published risk frameworks, run by the MIT AI Risk Initiative at MIT FutureTech, its unit \"the risk as a framework states it\" (MMRS record; its own pages).", "claims": [ "MMRS-01", "F2", "F44", "F34", "F35", "MMRS-03" ] }, "sections": [ { "icon": "🗂️", "head": "What it is", "items": [ { "label": "Three parts:", "text": "a database of 1700+ risks from 74 frameworks, a Causal Taxonomy and a Domain Taxonomy (repository page).", "claims": [ "F44", "F2" ] }, { "label": "Evidence:", "text": "each risk linked to its source, with quotes and page numbers (MIT IDE); frameworks published with excerpts (home page; MMRS record).", "claims": [ "F12", "F4", "MMRS-04" ] }, { "label": "What it gathers:", "text": "frameworks, themselves syntheses: \"A risk no framework has written down has no row\" (MMRS record); 65 documents from 17,000 screened (FAQ).", "claims": [ "MMRS-06", "MMRS-08", "F49" ] }, { "label": "Why:", "text": "no shared terminology for AI risks (arXiv v3).", "claims": [ "F25" ] } ] }, { "icon": "🧭", "head": "Taxonomies", "items": [ { "label": "Causal:", "text": "entity, intentionality, timing, each with \"Other\" (v1 abstract; repository page).", "claims": [ "F19", "F46" ] }, { "label": "Domain:", "text": "seven domains; 23, 24 or 25 subdomains by page; multi-agent risks added April 2025.", "claims": [ "F20", "F45", "F8", "F7", "MMRS-05" ] }, { "label": "Finding:", "text": "human decisions 38%, AI systems 42% (arXiv v3); the average framework covers 34% of subdomains (Slattery).", "claims": [ "F28", "F13" ] } ] }, { "icon": "📈", "head": "Versions", "items": [ { "label": "Counts:", "text": "777 from 43 (August 2024); 1,612 from 65 (April 2025); over 1,700 in Version 4 (December 2025), 74 frameworks in the paper's v3.", "claims": [ "F11", "F6", "F32", "F31", "F27" ] }, { "label": "Held apart:", "text": "the home page and the MMRS record say 65, the About page both 65+ and 70+; the Capture Registry calls 1,700 from 65 the current one.", "claims": [ "F3", "MMRS-13", "F37", "F38", "CAP-1-04" ] } ] }, { "icon": "⚠️", "head": "Limits", "items": [ { "label": "Stated:", "text": "one expert reviewer for extraction and coding; risks extracted as presented; no impact or likelihood (FAQ).", "claims": [ "F48", "F50" ] }, { "label": "Licence:", "text": "CC BY 4.0 (home page), which the MMRS record lists as unchecked.", "claims": [ "F5", "MMRS-17" ] } ] }, { "icon": "🔗", "head": "Around it", "items": [ { "label": "Incident Tracker:", "text": "an Initiative workstream with the AI Incident Database (About page); the MMRS record files it under the repository, a machine reading of AIID.", "claims": [ "F36", "MMRS-20", "MMRS-09" ] }, { "label": "Navigator:", "text": "\"a research prototype\", a distinct entity and a census candidate (its page; Capture Registry).", "claims": [ "F23", "CAP-1-05", "MMRS-23" ] }, { "label": "To machines:", "text": "its llms.txt asks assistants to check the pages before giving exact counts.", "claims": [ "F55", "F56" ] } ] } ], "open": [ { "question": "Is the framework count for version 4 65 or 74?", "rests_on": [ "F3", "F27", "F44", "F32", "F6", "MMRS-13", "CAP-1-02", "CAP-1-04" ], "modality": "documented on every side: the home page and the MMRS record 65, the paper's v3 and the repository's page 74; the Capture Registry's reading (interpretation) takes 65 as current", "would_resolve": "the Version 4 database's own document list, or a dated count on /risks" }, { "question": "How many subdomains has the Domain Taxonomy: 23, 24 or 25?", "rests_on": [ "F20", "F45", "F8" ], "modality": "stipulation (v1 abstract; repository page) against documented (April 2025 report post)", "would_resolve": "the taxonomy's own version history" }, { "question": "Is the AI Incident Tracker part of the repository or a sibling workstream of the Initiative?", "rests_on": [ "MMRS-20", "MMRS-14", "MMRS-02", "F36", "F41" ], "modality": "stipulation (MMRS record) against documented (About page; LinkedIn)", "would_resolve": "the Initiative's own statement of how its workstreams relate" }, { "question": "Can a risk enter without a framework that names it?", "rests_on": [ "MMRS-08", "MMRS-07", "F49", "F33" ], "modality": "interpretation (MMRS record) beside self-description (FAQ)", "would_resolve": "an intake path on the repository's pages for risks suggested by users or drawn from incidents, and whether suggestions enter as rows" }, { "question": "What does the ScienceDirect card's section on practical implications for technology management say?", "rests_on": [ "F43" ], "modality": "attributed (card snippet only; ScienceDirect and cell.com answered 403)", "would_resolve": "a fetch of the published article in Patterns (doi:10.1016/j.patter.2026.101517)" } ], "rail": [ { "lineage": "what it is: a living database of AI risks from frameworks", "claims": [ "MMRS-01", "F44", "F2", "F9", "F42", "F59" ], "earliest": "B2" }, { "lineage": "the unit and its evidence: a risk as a framework states it", "claims": [ "MMRS-03", "MMRS-12", "F12", "MMRS-04", "F4", "F33" ], "earliest": "B3" }, { "lineage": "purpose: a common frame of reference for AI risk", "claims": [ "F25", "F26", "F1", "F21", "F43" ], "earliest": "B4" }, { "lineage": "what it gathers: aggregates aggregators", "claims": [ "MMRS-06", "MMRS-07", "MMRS-08" ], "earliest": "MMRS" }, { "lineage": "stated limits", "claims": [ "F49", "F48", "F50", "F15" ], "earliest": "B3" }, { "lineage": "method: systematic review, expert consultation, best-fit synthesis", "claims": [ "F18", "F47", "F53" ], "earliest": "B4" }, { "lineage": "the Causal Taxonomy: entity, intent, timing", "claims": [ "F19", "F46" ], "earliest": "B4" }, { "lineage": "the Domain Taxonomy: seven domains and their subdomains", "claims": [ "F20", "MMRS-05", "F45", "F8", "F7" ], "earliest": "B4" }, { "lineage": "headline findings: coverage gaps and causes", "claims": [ "F28", "F14", "F13" ], "earliest": "B3" }, { "lineage": "count and version: 777/43 to 1,700+/74", "claims": [ "F11", "F17", "F6", "F31", "F32", "F27", "F3", "F37", "F38", "MMRS-13", "MMRS-15", "CAP-1-04" ], "earliest": "B4" }, { "lineage": "the Initiative's workstreams: the tracker beside the repository", "claims": [ "F41", "F36", "MMRS-14", "MMRS-20", "MMRS-02", "MMRS-21", "MMRS-09" ], "earliest": "B9" }, { "lineage": "the record kept apart from the entity", "claims": [ "MMRS-16", "MMRS-17", "MMRS-25" ], "earliest": "MMRS" }, { "lineage": "reception: the composition layer", "claims": [ "CAP-1-02", "MMRS-26", "CAP-1-01", "CAP-1-03" ], "earliest": "MMRS" }, { "lineage": "openness: free to copy, CC BY 4.0", "claims": [ "F5", "F40", "F51" ], "earliest": "B1" }, { "lineage": "the paper: Slattery et al., arXiv:2408.12622, Patterns 2026", "claims": [ "F30", "F29", "F10", "F54" ], "earliest": "B6" }, { "lineage": "operator: the MIT AI Risk Initiative at MIT FutureTech", "claims": [ "F34", "F35", "MMRS-11" ], "earliest": "B8" }, { "lineage": "place: other databases of AI risks, the harm basin", "claims": [ "MMRS-10", "MMRS-18", "MMRS-19", "MMRS-22", "F52" ], "earliest": "B11" }, { "lineage": "neighbour: the MIT AI Risk Navigator (research prototype)", "claims": [ "F24", "F22", "F23", "CAP-1-05", "CAP-1-06", "MMRS-23", "MMRS-24" ], "earliest": "B5" }, { "lineage": "machine-facing guidance: llms.txt", "claims": [ "F55", "F56", "F58", "F57" ], "earliest": "B12" }, { "lineage": "reach: users, citations, search", "claims": [ "F39", "F16" ], "earliest": "B3" } ], "body_words": 299, "entity": "Composed under v0.9, equal case (ρ = extends): KO_B sentences surviving 0 of 12; every field claim carried." }, "P_B": { "title": "MIT AI Risk Repository", "arm": "B (the field alone)", "genre": "popup — the compression from the field ledger alone (cards, own pages; no Wikipedia article)", "status": "composed 2026-10-29 from the field ledger alone; not frozen", "lede": { "text": "is the MIT AI Risk Initiative's \"living repository of AI risks\": a database of 1700+ risks from 74 frameworks with a Causal and a Domain Taxonomy (home page; repository page).", "claims": [ "F2", "F44", "F34" ] }, "sections": [ { "icon": "🗂️", "head": "What it is", "items": [ { "label": "Evidence:", "text": "each risk linked to its source, quotes and page numbers (MIT IDE).", "claims": [ "F12" ] }, { "label": "Method:", "text": "systematic review, expert consultation, best-fit synthesis (v1 abstract).", "claims": [ "F18" ] } ] }, { "icon": "🧭", "head": "Taxonomies", "items": [ { "label": "Causal:", "text": "entity, intent, timing (v1 abstract).", "claims": [ "F19" ] }, { "label": "Domain:", "text": "seven domains, 24 subdomains (repository page).", "claims": [ "F20", "F45" ] }, { "label": "Finding:", "text": "human decisions 38%, AI systems 42% (arXiv v3).", "claims": [ "F28" ] } ] }, { "icon": "📈", "head": "Versions", "items": [ { "label": "Counts:", "text": "777 from 43 (2024) to over 1,700 in Version 4 (December 2025).", "claims": [ "F11", "F32", "F31" ] } ] }, { "icon": "⚠️", "head": "Limits", "items": [ { "label": "Stated:", "text": "a single expert reviewer; risks extracted as presented (FAQ).", "claims": [ "F48" ] } ] } ], "open": [ { "question": "65 or 74 frameworks?", "rests_on": [ "F3", "F27", "F44" ], "modality": "documented, both", "would_resolve": "a dated count on /risks" } ], "rail": [ { "lineage": "what it is: a living database of AI risks from frameworks", "claims": [ "F2", "F44", "F9", "F42", "F59" ], "earliest": "B2" }, { "lineage": "purpose: a common frame of reference for AI risk", "claims": [ "F25", "F26", "F1", "F21", "F43" ], "earliest": "B4" }, { "lineage": "the unit and its evidence: a risk as a framework states it", "claims": [ "F12", "F4", "F33" ], "earliest": "B3" }, { "lineage": "method: systematic review, expert consultation, best-fit synthesis", "claims": [ "F18", "F47", "F53" ], "earliest": "B4" }, { "lineage": "the Causal Taxonomy: entity, intent, timing", "claims": [ "F19", "F46" ], "earliest": "B4" }, { "lineage": "the Domain Taxonomy: seven domains and their subdomains", "claims": [ "F20", "F45", "F8", "F7" ], "earliest": "B4" }, { "lineage": "headline findings: coverage gaps and causes", "claims": [ "F28", "F14", "F13" ], "earliest": "B3" }, { "lineage": "count and version: 777/43 to 1,700+/74", "claims": [ "F11", "F17", "F6", "F31", "F32", "F27", "F3", "F37", "F38" ], "earliest": "B4" }, { "lineage": "the Initiative's workstreams: the tracker beside the repository", "claims": [ "F41", "F36" ], "earliest": "B9" }, { "lineage": "stated limits", "claims": [ "F48", "F49", "F50", "F15" ], "earliest": "B3" }, { "lineage": "operator: the MIT AI Risk Initiative at MIT FutureTech", "claims": [ "F34", "F35" ], "earliest": "B8" }, { "lineage": "the paper: Slattery et al., arXiv:2408.12622, Patterns 2026", "claims": [ "F29", "F10", "F30", "F54" ], "earliest": "B6" }, { "lineage": "openness: free to copy, CC BY 4.0", "claims": [ "F5", "F40", "F51" ], "earliest": "B1" }, { "lineage": "neighbour: the MIT AI Risk Navigator (research prototype)", "claims": [ "F22", "F23", "F24" ], "earliest": "B5" }, { "lineage": "place: other databases of AI risks, the harm basin", "claims": [ "F52" ], "earliest": "B11" }, { "lineage": "machine-facing guidance: llms.txt", "claims": [ "F55", "F56", "F57", "F58" ], "earliest": "B12" }, { "lineage": "reach: users, citations, search", "claims": [ "F39", "F16" ], "earliest": "B3" } ], "body_words": 98, "entity": "With no archive claim admitted, the entry is the field's database at its own resolution: its pages, its paper and its counts by version, its limits stated, the tracker and the Navigator beside it." }, "KO_B": { "title": "The MIT AI Risk Repository", "arm": "B (the field alone)", "genre": "knowledge object", "status": "composed 2026-10-29; not frozen", "sentences": [ { "n": 1, "para": 0, "text": "The MIT AI Risk Repository is the MIT AI Risk Initiative's living repository of AI risks, which its page describes in three parts, an AI Risk Database of 1700+ risks from 74 frameworks, a Causal Taxonomy and a Domain Taxonomy, a structured and evolving resource across literatures; LinkedIn's page counts over 1000 risks, and the file for machines points to /risks.", "claims": [ "F2", "F44", "F9", "F42", "F59" ] }, { "n": 2, "para": 0, "text": "It answers a lack of shared terminology for AI risks, privacy meaning leaked training data in one framework and freedom from surveillance in another, with authoritative data and frameworks for identifying and managing risks, the first such categorized risk database by its authors' account, and tools for managing AI's societal impact.", "claims": [ "F25", "F26", "F1", "F21", "F43" ] }, { "n": 3, "para": 1, "text": "Each risk carries its source and supporting evidence, quotes and page numbers, and the frameworks are published in a deck with their documents and excerpts; a framework added in Version 4 was built from 499 reported incidents.", "claims": [ "F12", "F4", "F33" ] }, { "n": 4, "para": 1, "text": "It was made by systematic review and expert consultation, its taxonomies by best-fit framework synthesis, because causal and hazard categories would not map onto one another.", "claims": [ "F18", "F47", "F53" ] }, { "n": 5, "para": 2, "text": "The Causal Taxonomy codes entity, intentionality and timing, each with a value Other.", "claims": [ "F19", "F46" ] }, { "n": 6, "para": 2, "text": "The Domain Taxonomy has seven domains with 23, 24 or 25 subdomains by page, multi-agent risks added in April 2025.", "claims": [ "F20", "F45", "F8", "F7" ] }, { "n": 7, "para": 3, "text": "In v3 human decisions account for 38% of risks and AI systems for 42%; the first release had 51% and 34%, and the average framework covered 34% of subdomains.", "claims": [ "F28", "F14", "F13" ] }, { "n": 8, "para": 3, "text": "Its size has grown from 777 risks and 43 frameworks in August 2024 to 1,612 and 65 in April 2025 and over 1,700 in Version 4 of December 2025, 74 frameworks in v3 of the paper, while the home page says 65, the About page 65 and 70, and LinkedIn 1,700.", "claims": [ "F11", "F17", "F6", "F31", "F32", "F27", "F3", "F37", "F38", "F41" ] }, { "n": 9, "para": 4, "text": "Its limits are a single expert reviewer, risks extracted as presented, 65 documents that may miss emerging and unpublished risks, no impact or likelihood, and the need for modification before most audiences can use it.", "claims": [ "F48", "F49", "F50", "F15" ] }, { "n": 10, "para": 4, "text": "The MIT AI Risk Initiative, based at MIT FutureTech, runs it; the paper by Slattery and nine others dates from August 2024, was revised in April 2025 and May 2026 and published in Patterns; the data is CC BY 4.0 and free to copy.", "claims": [ "F34", "F35", "F29", "F10", "F30", "F54", "F5", "F40", "F51" ] }, { "n": 11, "para": 5, "text": "Beside it the Initiative runs an Incident Tracker with the AI Incident Database and the Responsible AI Collaborative, and a Navigator that is a research prototype by Spencer Michaels; the repository names MITRE ATT&CK, airisk.io, AVID and aitracker.org as other databases.", "claims": [ "F36", "F41", "F22", "F23", "F24", "F52" ] }, { "n": 12, "para": 5, "text": "Its file for language models asks them to name the Initiative, check counts before stating them, tell incidents from risks and give no legal advice, and its About page counts 290,000 users and calls the site a top search result on AI risks.", "claims": [ "F55", "F56", "F57", "F58", "F39", "F16" ] } ] }, "L_B": { "words": 459, "text": "The MIT AI Risk Repository is the MIT AI Risk Initiative's living repository of AI risks, which its page describes in three parts, an AI Risk Database of 1700+ risks from 74 frameworks, a Causal Taxonomy and a Domain Taxonomy, a structured and evolving resource across literatures; LinkedIn's page counts over 1000 risks, and the file for machines points to /risks. It answers a lack of shared terminology for AI risks, privacy meaning leaked training data in one framework and freedom from surveillance in another, with authoritative data and frameworks for identifying and managing risks, the first such categorized risk database by its authors' account, and tools for managing AI's societal impact. Each risk carries its source and supporting evidence, quotes and page numbers, and the frameworks are published in a deck with their documents and excerpts; a framework added in Version 4 was built from 499 reported incidents. It was made by systematic review and expert consultation, its taxonomies by best-fit framework synthesis, because causal and hazard categories would not map onto one another. The Causal Taxonomy codes entity, intentionality and timing, each with a value Other. The Domain Taxonomy has seven domains with 23, 24 or 25 subdomains by page, multi-agent risks added in April 2025. In v3 human decisions account for 38% of risks and AI systems for 42%; the first release had 51% and 34%, and the average framework covered 34% of subdomains. Its size has grown from 777 risks and 43 frameworks in August 2024 to 1,612 and 65 in April 2025 and over 1,700 in Version 4 of December 2025, 74 frameworks in v3 of the paper, while the home page says 65, the About page 65 and 70, and LinkedIn 1,700. Its limits are a single expert reviewer, risks extracted as presented, 65 documents that may miss emerging and unpublished risks, no impact or likelihood, and the need for modification before most audiences can use it. The MIT AI Risk Initiative, based at MIT FutureTech, runs it; the paper by Slattery and nine others dates from August 2024, was revised in April 2025 and May 2026 and published in Patterns; the data is CC BY 4.0 and free to copy. Beside it the Initiative runs an Incident Tracker with the AI Incident Database and the Responsible AI Collaborative, and a Navigator that is a research prototype by Spencer Michaels; the repository names MITRE ATT&CK, airisk.io, AVID and aitracker.org as other databases. Its file for language models asks them to name the Initiative, check counts before stating them, tell incidents from risks and give no legal advice, and its About page counts 290,000 users and calls the site a top search result on AI risks.", "note": "L_F(B) is realized as the field arm's expansion (objects.KO_B): the field alone, cards and own pages (no Wikipedia article)." }, "L_BA": { "words": 1703, "text": "The MIT AI Risk Repository is a database of more than 1,700 AI risks extracted from published risk frameworks, run by the MIT AI Risk Initiative (MIT FutureTech) (MMRS record); its own page gives it three parts, an AI Risk Database that \"captures 1700+ risks extracted from 74 existing frameworks and classifications of AI risks\", a Causal Taxonomy and a Domain Taxonomy, and the Initiative's home page calls it a \"living repository of AI risks\". Its unit is the risk as a framework states it (MMRS record), \"risk, extracted from published risk frameworks\" in the record's field, each risk linked to \"the source information (paper title, authors) and supporting evidence, such as quotes and page numbers\" (MIT IDE), and each source framework summarized with its source documents and key excerpts published (MMRS record), as the home page's deck of the 65 Version 3 frameworks gives them. Its report post of April 2025 calls it \"a structured and evolving resource\" consolidating how AI risks are categorized across academic, industry and policy literatures, its LinkedIn page \"a comprehensive living database of over 1000 AI risks categorized by their cause and risk domain\", and the Initiative's file for machines names /risks its canonical page, with \"methodology notes, use cases, limitations, data access links, and citation guidance\". Slattery and colleagues start from terminology: \"researchers, policymakers, and technology companies lack shared terminology for discussing AI risks\", so that \"privacy\" names a model's ability to leak sensitive training data in one framework and freedom from government surveillance in another (arXiv:2408.12622, v3). The Initiative offers \"Authoritative data and frameworks to help you identify, prioritize, and manage the risks from AI\" (home page); the v1 abstract calls the repository, to its authors' knowledge, \"the first attempt to rigorously curate, analyze, and extract AI risk frameworks into a publicly accessible, comprehensive, extensible, and categorized risk database\", and the published article's snippet promises \"concrete tools for managing AI's societal impact across multiple stakeholder[s]\" (ScienceDirect card). The MMRS record reads the same object as an aggregator of aggregators, \"its sources are frameworks, which are themselves syntheses\": \"What enters is what a framework named\", and \"A risk no framework has written down has no row\"; the repository's FAQ states the boundary as its limit, 65 documents from more than 17,000 records screened, which \"May be missing emerging, domain-specific risks, and unpublished risks\", and a framework can itself reach back to events, as the taxonomy of generative AI failures built from 499 publicly reported incidents that Version 4 added (December 2025 update). Its method is \"a systematic review of taxonomies and other structured classifications of AI risk followed by an expert consultation\", its taxonomies made by \"a best-fit framework synthesis\" (v1 abstract), restated in the FAQ as a systematic search with forwards and backwards searching and expert consultation that identified 65 documents and extracted 1600+ risks; there are two taxonomies because causal categories such as \"Diffusion of responsibility\" do not map to hazard categories such as \"Misuse\" (FAQ). The Causal Taxonomy classifies each risk by Entity (Human, AI), Intentionality (Intentional, Unintentional) and Timing (Pre-deployment, Post-deployment) (v1 abstract), and the repository's page gives each factor a third value, \"Other: Due to some other reason or ambiguous\" for entity. The Domain Taxonomy sorts risks into seven domains, discrimination and toxicity, privacy and security, misinformation, malicious actors and misuse, human-computer interaction, socioeconomic and environmental, and AI system safety, failures and limitations (v1 abstract), each with subdomains (MMRS record): 23 in v1, 24 on the repository's page, 25 in the April 2025 report post, the subdomain for multi-agent risks added in April 2025 (MMRS record; report post). Its v3 finding is that, \"contrary to common assumptions, human decisions cause nearly as many AI risks (38%) as the AI systems themselves (42%)\" (arXiv v3); the first release had found 51% of risks caused by AI systems and 34% by humans, 65% after deployment and 10% before, and Slattery gave as its key finding \"the average framework covering only 34% of the identified risk subdomains and even the most significant frameworks covering only 70%\" (MIT IDE). The counts are versioned: 777 risks from 43 frameworks at the first release, \"A new MIT FutureTech project\" (MIT IDE, 21 August 2024), \"777 risks extracted from 43 taxonomies\" in the v1 abstract; 22 frameworks added to 65 documents and 1,612 unique risk entries in April 2025 (report post); Version 4 live on 4 December 2025 with 9 frameworks and about 200 risk categories added, \"over 1,700 coded risks\" (December 2025 update); and \"74 frameworks containing 1,725 distinct risks\" in v3 of the paper, repeated on the repository's page as 1700+ from 74. Older figures stand beside the current one on the Initiative's own pages: \"1,700+ AI risks from 65 frameworks\" on the home page, \"more than 1,600 risks from more than 65 frameworks and papers\" and, further down the same About page, \"1700+ risks from 70+ frameworks\", and \"more than 1,700\" in the LinkedIn post; the MMRS record gives \"more than 1,700 risks from 65 frameworks\" as of 2026-10-28, read from airisk.mit.edu for \"scope, version\", with lifecycle \"version 4, December 2025\" (MMRS record). The Capture Registry takes 1,700 from 65 as the record's count for version 4 and finds that, of the three sizes on the cards, \"the composition gives the current one\" (Capture Registry, 2026-10-29); version 4's post adds 9 frameworks to version 3's 65, and v3 of the paper and the repository's page give 74 while the home page gives 65, so the record's 65 is held beside the 74, each with its source. Its stated limits are \"a single expert reviewer for extraction and coding\", risks extracted as presented, poorly communicated ones included, and taxonomies that \"prioritize clarity and simplicity over nuance\" and do not categorize risks by impact, likelihood or their interaction (repository FAQ); Slattery expected it to \"need some modification before it is useful for most audiences\" (MIT IDE). The MMRS record lists extraction method and licence as unchecked; the field states both, the method above and \"Data from the MIT AI Risk Initiative is licensed under CC BY 4.0\" (home page), \"most of our outputs are published under a CC BY 4.0 licence\" (About page), \"Our Database is free to copy and use\" (repository page), the paper CC BY 4.0 in Patterns (arXiv). Its operator is the MIT AI Risk Initiative (AIRI), \"based at MIT FutureTech, a joint initiative of the Institute of the Digital Economy and MIT CSAIL\" (About page), as the MMRS record gives it, on the operator's statement. Its paper is by Peter Slattery, Alexander K. Saeri, Emily A. C. Grundy, Jess Graham, Michael Noetel, Risto Uuk, James Dao, Soroush Pour, Stephen Casper and Neil Thompson, submitted on 14 August 2024 and revised to v3 on 5 May 2026 (arXiv), with versions of August 2024 and April 2025 (report post), published in Patterns (Cell Press, 2026), and the repository asks to be cited by its preprint (FAQ). The repository is one of the Initiative's \"several connected workstreams, each answering a different question\", beside an AI Incident Tracker \"Developed in collaboration with the AI Incident Database\" (About page), which the Initiative's LinkedIn post calls \"a collaboration with the Responsible AI Collaborative, which connects risks to over 1,400 incidents\". The MMRS record keeps the incidents apart from the risks, \"real-world incidents are kept separately in its AI Incident Tracker\", records the tracker as part of the repository, classifying AI Incident Database reports against the same taxonomy through Arcola AI's language-model pipeline, and reads that layer as \"a second-order reading of the AI Incident Database, done by machine\" (MMRS record); the field's About page sets the tracker beside the repository as a workstream, and the two placements stand each with its source. In the record's harm basin the repository \"sits over the harm basin as a taxonomy for it\", its role harm and its neighbour the AI Incident Database, whose own record says the repository \"aggregates risk frameworks over the same ground and keeps its own incident tracker\" (MMRS record); the repository's FAQ names as other databases of AI risks MITRE ATT&CK, airisk.io, AVID and aitracker.org. The MIT AI Risk Navigator, \"A project of the AI Risk Initiative at MIT FutureTech\" with the support of the Cambridge Boston Alignment Initiative, connects the repository's datasets \"through a shared taxonomy\" and says of itself \"This is a research prototype. The data and analyses are preliminary and not yet validated\" (its home page); the Capture Registry finds it \"a third MIT surface\" while \"the census holds the repository and the tracker\", \"Added as a candidate\", and the MMRS record lists it as a candidate to enter \"once it has been read at its own pages\", a distinct entity. The Initiative publishes a file \"to orient LLMs and AI agents\" that asks them to \"Refer to the project as the MIT AI Risk Initiative\", to \"Check the linked pages before giving exact counts, dates, version numbers, or coverage claims\", to \"distinguish reported incidents from predicted risks\", to note that the Navigator \"is a research prototype\", and not to present the datasets \"as legal, safety, or investment advice\" (llms.txt). Its About page counts 250+ citations, the 2025 and 2026 International AI Safety Reports among them, and 290,000+ website users, and calls the site \"One of the top sources returned for search queries on AI risks and incidents\"; in its first week over 35,000 people used the website (MIT IDE). The MMRS record has \"two nodes, the MMRS record and the entity it describes\", each assertion with a source and a basis and the lifecycle \"kept apart from the observation that saw it\", and it links the capture as \"the composition layer's account of the entity, seated in the Capture Registry\" (MMRS record). At the entity's name the Capture Registry finds \"The repository is composed at a finer grain than the record holds, its taxonomy named domain by domain\", the seven domains and the causal taxonomy \"the record's 'domain taxonomy with subdomains' given at full grain\", checked against the record and the entity's own pages on 2026-10-29 (Capture Registry, 2026-10-29).", "rail": [], "note": "L(B ∪ A) in the equal-terms case (v0.9 §0.4), realized as the knowledge object (objects.KO); its rail is the compression's, by lineage." } }, "field": [ { "id": "B1", "card": "MIT AI Risk Initiative — MIT AI Risk Initiative", "fetched": "card; HTTP 200; B1.html sha256 d3f35366280a66ca", "origin": "card" }, { "id": "B2", "card": "MIT AI Risk Initiative — AI Risk Repository Report updated (April 2025)", "fetched": "card; HTTP 200; B2.html sha256 437d6f93ad588640", "origin": "card" }, { "id": "B3", "card": "MIT Initiative on the Digital Economy — MIT Researchers Create an AI Risk Repository - MIT Initiative on the Digital Economy", "fetched": "card; HTTP 200; B3.html sha256 75803d35fbe47a7c", "origin": "card" }, { "id": "B4", "card": "MIT FutureTech — FutureTech", "fetched": "card; HTTP 200; B4.html sha256 d8b51fa0df9e925a", "origin": "card" }, { "id": "B5", "card": "MIT AI Risk Navigator — MIT AI Risk Navigator", "fetched": "card; HTTP 200; B5.html sha256 aff8629414802228", "origin": "card" }, { "id": "B6", "card": "arXiv — [2408.12622] The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence", "fetched": "card; HTTP 200; B6.html sha256 79eb6de30ec0f027", "origin": "card" }, { "id": "B7", "card": "MIT AI Risk Initiative — Repository Update: December 2025", "fetched": "card; HTTP 200; B7.html sha256 8afe267cf8dd329a", "origin": "card" }, { "id": "B8", "card": "MIT AI Risk Initiative — About", "fetched": "card; HTTP 200; B8.html sha256 b40b72f7dbd5c16c", "origin": "card" }, { "id": "B9", "card": "LinkedIn — MIT AI Risk Initiative | LinkedIn", "fetched": "card; HTTP 200; B9.html sha256 d098a0edbdf235f7", "origin": "card" }, { "id": "B10", "card": "ScienceDirect.com — The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence - ScienceDirect", "fetched": "card; HTTP 403; B10.html sha256 c307c5915d333b3f (403 body; snippet stands)", "origin": "card" }, { "id": "B11", "card": "own page: the repository's own page (AI Risk Repository, airisk.mit.edu/risks)", "fetched": "own_page; HTTP 200; B11.html sha256 a5bd76e2fcecfd10", "origin": "own_page" }, { "id": "B12", "card": "own page: airisk.mit.edu/llms.txt, the Initiative's file for language models and AI agents", "fetched": "own_page; HTTP 200; B12.html sha256 039e65007ed87700", "origin": "own_page" }, { "id": "W1", "card": "no article", "fetched": "wikipedia; HTTP 404 (raw and rendered); Special:Search one mention", "origin": "wikipedia" } ], "field_claims": { "F1": { "source": "B1", "origin": "card", "locus": "hero ¶1–2", "quote": "Producing the information needed to address risks from AI … Authoritative data and frameworks to help you identify, prioritize, and manage the risks from AI", "claim": "the Initiative's home page: authoritative data and frameworks to identify, prioritize and manage AI risks", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "purpose: a common frame of reference for AI risk", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 935d8b0170045fd4…), whitespace-normalised" }, "F2": { "source": "B1", "origin": "card", "locus": "Our tools and data: 'What are the risks from AI?'", "quote": "Learn about the many different AI risks and how they emerge with our living repository of AI risks", "claim": "the repository named on the home page as the Initiative's living repository of AI risks", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "what it is: a living database of AI risks from frameworks", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 935d8b0170045fd4…), whitespace-normalised" }, "F3": { "source": "B1", "origin": "card", "locus": "All collections: AI Risks", "quote": "A living database of 1,700+ AI risks from 65 frameworks", "claim": "the home page's count: 1,700+ risks from 65 frameworks", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 935d8b0170045fd4…), whitespace-normalised" }, "F4": { "source": "B1", "origin": "card", "locus": "blog: 'Explore the Frameworks Behind the AI Risk Repository'", "quote": "This post shares a public Google Slides deck containing all 65 frameworks included in Version 3 of the AI Risk Repository. It provides direct access to source documents, diagrams, and key excerpts to support transparency, research, and further analysis of AI risk classifications.", "claim": "the 65 Version 3 frameworks published as a deck with source documents, diagrams and key excerpts", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the unit and its evidence: a risk as a framework states it", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 935d8b0170045fd4…), whitespace-normalised" }, "F5": { "source": "B1", "origin": "card", "locus": "footer", "quote": "Data from the MIT AI Risk Initiative is licensed under CC BY 4.0", "claim": "the Initiative's data licensed CC BY 4.0", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "openness: free to copy, CC BY 4.0", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 935d8b0170045fd4…), whitespace-normalised" }, "F6": { "source": "B2", "origin": "card", "locus": "What's new in this update", "quote": "Integration of 22 new AI risk frameworks, bringing the total number of included documents to 65. … Expansion of the AI Risk Database to 1,612 unique risk entries, each systematically extracted and coded.", "claim": "the April 2025 update: 22 new frameworks to 65 documents, 1,612 unique risk entries", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2025-04-23 (post date)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c6e93a76c8af929f…), whitespace-normalised" }, "F7": { "source": "B2", "origin": "card", "locus": "What's new in this update", "quote": "Introduction of a new risk subdomain on multi-agent risks, reflecting recent developments in AI research on complex agent interactions.", "claim": "a new subdomain on multi-agent risks", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the Domain Taxonomy: seven domains and their subdomains", "date": "2025-04-23 (post date)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c6e93a76c8af929f…), whitespace-normalised" }, "F8": { "source": "B2", "origin": "card", "locus": "About the AI Risk Repository", "quote": "A Domain Taxonomy, which classifies risks into seven thematic domains and 25 more granular subdomains.", "claim": "the domain taxonomy given as seven domains and 25 subdomains (other own pages give 24)", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the Domain Taxonomy: seven domains and their subdomains", "date": "2025-04-23 (post date)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c6e93a76c8af929f…), whitespace-normalised" }, "F9": { "source": "B2", "origin": "card", "locus": "About the AI Risk Repository ¶1", "quote": "The AI Risk Repository is a structured and evolving resource that aims to consolidate and clarify how risks from artificial intelligence are categorized across academic, industry, and policy literatures.", "claim": "a structured, evolving resource consolidating how AI risks are categorized across literatures", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "what it is: a living database of AI risks from frameworks", "date": "2025-04-23 (post date)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c6e93a76c8af929f…), whitespace-normalised" }, "F10": { "source": "B2", "origin": "card", "locus": "header", "quote": "Updated Version of the AI Risk Repository Preprint Now Available … Original release: August 2024 … Updated version: April 2025", "claim": "the preprint's versions: August 2024 original, April 2025 update", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the paper: Slattery et al., arXiv:2408.12622, Patterns 2026", "date": "2025-04-23 (post date)", "quote_ok": true, "checked_against": "saved text B2.txt (sha256 c6e93a76c8af929f…), whitespace-normalised" }, "F11": { "source": "B3", "origin": "card", "locus": "lede ¶1", "quote": "A new MIT FutureTech project reviewed 43 AI frameworks produced by research, industry and government organizations; they identified 777 risks in total.", "claim": "the first release: an MIT FutureTech project, 43 frameworks, 777 risks", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2024-08-21 (post date)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 e3c9288651eb1a00…), whitespace-normalised" }, "F12": { "source": "B3", "origin": "card", "locus": "lede ¶2", "quote": "The repository includes a risk database linking each risk to the source information (paper title, authors) and supporting evidence, such as quotes and page numbers.", "claim": "each risk linked to its source and supporting evidence (quotes, page numbers)", "modality": "documented", "kind": "function", "kernel": true, "const": false, "lineage": "the unit and its evidence: a risk as a framework states it", "date": "2024-08-21 (post date)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 e3c9288651eb1a00…), whitespace-normalised" }, "F13": { "source": "B3", "origin": "card", "locus": "interview: key findings", "quote": "The key finding from our analysis is that there are significant gaps in existing risk frameworks, with the average framework covering only 34% of the identified risk subdomains and even the most significant frameworks covering only 70%.", "claim": "Slattery: the average framework covers 34% of subdomains, the most comprehensive 70%", "modality": "attributed", "kind": "interpretation", "kernel": true, "const": false, "lineage": "headline findings: coverage gaps and causes", "date": "2024-08-21 (post date)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 e3c9288651eb1a00…), whitespace-normalised" }, "F14": { "source": "B3", "origin": "card", "locus": "interview: key findings", "quote": "most of the risks (51%) were caused by AI systems rather than humans (34%), and were found after the AI model was trained and deployed (65%) rather than before (10%).", "claim": "the first release's causal figures: AI 51%, human 34%; post-deployment 65%, pre-deployment 10%", "modality": "attributed", "kind": "fact", "kernel": false, "const": false, "lineage": "headline findings: coverage gaps and causes", "date": "2024-08-21 (post date)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 e3c9288651eb1a00…), whitespace-normalised" }, "F15": { "source": "B3", "origin": "card", "locus": "interview: audience", "quote": "We expect that what we have produced will need some modification before it is useful for most audiences, but we hope that it provides a solid foundation.", "claim": "Slattery: it will need modification before it is useful for most audiences", "modality": "attributed", "kind": "limit", "kernel": false, "const": false, "lineage": "stated limits", "date": "2024-08-21 (post date)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 e3c9288651eb1a00…), whitespace-normalised" }, "F16": { "source": "B3", "origin": "card", "locus": "interview: reception", "quote": "In less than a week, over 35,000 people have used the website and over 6,000 have viewed our explainer video on YouTube.", "claim": "first week: over 35,000 website users, over 6,000 views of the explainer", "modality": "attributed", "kind": "fact", "kernel": false, "const": false, "lineage": "reach: users, citations, search", "date": "2024-08-21 (post date)", "quote_ok": true, "checked_against": "saved text B3.txt (sha256 e3c9288651eb1a00…), whitespace-normalised" }, "F17": { "source": "B4", "origin": "card", "locus": "abstract", "quote": "This comprises a living database of 777 risks extracted from 43 taxonomies, which can be filtered based on two overarching taxonomies and easily accessed, modified, and updated via our website and online spreadsheets.", "claim": "v1: a living database of 777 risks from 43 taxonomies, filtered by two taxonomies, via website and spreadsheets", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2024-08-14", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 56497bd8702d8e73…), whitespace-normalised" }, "F18": { "source": "B4", "origin": "card", "locus": "abstract", "quote": "We construct our Repository with a systematic review of taxonomies and other structured classifications of AI risk followed by an expert consultation. We develop our taxonomies of AI risk using a best-fit framework synthesis.", "claim": "method: systematic review, then expert consultation; taxonomies by best-fit framework synthesis", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "method: systematic review, expert consultation, best-fit synthesis", "date": "2024-08-14", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 56497bd8702d8e73…), whitespace-normalised" }, "F19": { "source": "B4", "origin": "card", "locus": "abstract", "quote": "Our high-level Causal Taxonomy of AI Risks classifies each risk by its causal factors (1) Entity: Human, AI; (2) Intentionality: Intentional, Unintentional; and (3) Timing: Pre-deployment; Post-deployment.", "claim": "the Causal Taxonomy: entity, intentionality, timing", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "the Causal Taxonomy: entity, intent, timing", "date": "2024-08-14", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 56497bd8702d8e73…), whitespace-normalised" }, "F20": { "source": "B4", "origin": "card", "locus": "abstract", "quote": "Our mid-level Domain Taxonomy of AI Risks classifies risks into seven AI risk domains: (1) Discrimination & toxicity, (2) Privacy & security, (3) Misinformation, (4) Malicious actors & misuse, (5) Human-computer interaction, (6) Socioeconomic & environmental, and (7) AI system safety, failures, & limitations. These are further divided into 23 subdomains.", "claim": "the Domain Taxonomy: seven domains named, 23 subdomains (v1)", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "the Domain Taxonomy: seven domains and their subdomains", "date": "2024-08-14", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 56497bd8702d8e73…), whitespace-normalised" }, "F21": { "source": "B4", "origin": "card", "locus": "abstract", "quote": "The AI Risk Repository is, to our knowledge, the first attempt to rigorously curate, analyze, and extract AI risk frameworks into a publicly accessible, comprehensive, extensible, and categorized risk database.", "claim": "to its authors' knowledge the first such curated, extracted, categorized risk database", "modality": "self-description", "kind": "interpretation", "kernel": false, "const": false, "lineage": "purpose: a common frame of reference for AI risk", "date": "2024-08-14", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 56497bd8702d8e73…), whitespace-normalised" }, "F22": { "source": "B5", "origin": "card", "locus": "hero", "quote": "The Navigator connects MIT's AI Risk Repository datasets through a shared taxonomy, so you can move between them and surface the patterns that matter.", "claim": "the Navigator connects the datasets through a shared taxonomy", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "neighbour: the MIT AI Risk Navigator (research prototype)", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 0b1d110533f1a4e8…), whitespace-normalised" }, "F23": { "source": "B5", "origin": "card", "locus": "banner", "quote": "This is a research prototype. The data and analyses are preliminary and not yet validated", "claim": "the Navigator's limit: a research prototype, preliminary, not yet validated", "modality": "self-description", "kind": "limit", "kernel": true, "const": false, "lineage": "neighbour: the MIT AI Risk Navigator (research prototype)", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 0b1d110533f1a4e8…), whitespace-normalised" }, "F24": { "source": "B5", "origin": "card", "locus": "hero ¶2; footer", "quote": "A project of the AI Risk Initiative at MIT FutureTech, with the generous support of the Cambridge Boston Alignment Initiative. … © 2026 Spencer Michaels", "claim": "the Navigator a project of the AI Risk Initiative, supported by the Cambridge Boston Alignment Initiative; site by Spencer Michaels", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "neighbour: the MIT AI Risk Navigator (research prototype)", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 0b1d110533f1a4e8…), whitespace-normalised" }, "F25": { "source": "B6", "origin": "card", "locus": "Abstract", "quote": "Yet researchers, policymakers, and technology companies lack shared terminology for discussing AI risks.", "claim": "the problem: no shared terminology for AI risks", "modality": "interpretation", "kind": "interpretation", "kernel": true, "const": false, "lineage": "purpose: a common frame of reference for AI risk", "date": "2026-05-05 (v3)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 7726a68799cf5a26…), whitespace-normalised" }, "F26": { "source": "B6", "origin": "card", "locus": "Abstract", "quote": "Consider \"privacy\": one framework uses this term to describe a model's ability to leak sensitive training data, while another uses it to mean freedom from government surveillance.", "claim": "the example: 'privacy' in two frameworks names two different things", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "purpose: a common frame of reference for AI risk", "date": "2026-05-05 (v3)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 7726a68799cf5a26…), whitespace-normalised" }, "F27": { "source": "B6", "origin": "card", "locus": "Abstract", "quote": "We systematically analyzed every major AI risk framework published to date-74 frameworks containing 1,725 distinct risks-and organized them into a unified system.", "claim": "v3: 74 frameworks containing 1,725 distinct risks, organized into one system", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2026-05-05 (v3)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 7726a68799cf5a26…), whitespace-normalised" }, "F28": { "source": "B6", "origin": "card", "locus": "Abstract", "quote": "Our two classification systems reveal important patterns: contrary to common assumptions, human decisions cause nearly as many AI risks (38%) as the AI systems themselves (42%).", "claim": "v3 finding: human decisions 38%, AI systems 42%", "modality": "documented", "kind": "interpretation", "kernel": true, "const": false, "lineage": "headline findings: coverage gaps and causes", "date": "2026-05-05 (v3)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 7726a68799cf5a26…), whitespace-normalised" }, "F29": { "source": "B6", "origin": "card", "locus": "header; Authors", "quote": "[Submitted on 14 Aug 2024 (v1), last revised 5 May 2026 (this version, v3)] … Authors:Peter Slattery, Alexander K. Saeri, Emily A. C. Grundy, Jess Graham, Michael Noetel, Risto Uuk, James Dao, Soroush Pour, Stephen Casper, Neil Thompson", "claim": "the paper's dates (v1 14 Aug 2024, v3 5 May 2026) and its ten authors, Slattery first, Thompson last", "modality": "documented", "kind": "genealogy", "kernel": true, "const": false, "lineage": "the paper: Slattery et al., arXiv:2408.12622, Patterns 2026", "date": "2024-08-14 (v1)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 7726a68799cf5a26…), whitespace-normalised" }, "F30": { "source": "B6", "origin": "card", "locus": "Comments; Journal reference", "quote": "This paper has been published in Patterns (Cell Press, 2026) under a CC BY 4.0 licence … Patterns 101517 (2026)", "claim": "published in Patterns (Cell Press, 2026), CC BY 4.0", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the paper: Slattery et al., arXiv:2408.12622, Patterns 2026", "date": "2026-05-05 (v3)", "quote_ok": true, "checked_against": "saved text B6.txt (sha256 7726a68799cf5a26…), whitespace-normalised" }, "F31": { "source": "B7", "origin": "card", "locus": "¶1", "quote": "We’re pleased to share that Version 4 of the MIT AI Risk Repository is now live.", "claim": "Version 4 live (4 December 2025)", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2025-12-04 (post date)", "quote_ok": true, "checked_against": "saved text B7.txt (sha256 21849cbdbd3f0259…), whitespace-normalised" }, "F32": { "source": "B7", "origin": "card", "locus": "What's new in Version 4", "quote": "9 newly added frameworks, spanning government & industry reports, peer reviewed journal articles, conference papers and preprints, … ~200 new AI risk categories, expanding the repository to over 1,700 coded risks", "claim": "Version 4: 9 frameworks added, ~200 risk categories, over 1,700 coded risks", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2025-12-04 (post date)", "quote_ok": true, "checked_against": "saved text B7.txt (sha256 21849cbdbd3f0259…), whitespace-normalised" }, "F33": { "source": "B7", "origin": "card", "locus": "Newly added frameworks: Li et al., 2025", "quote": "This preprint presents a taxonomy of Generative AI Failures based on a systematic analysis of 499 publicly reported AI incidents.", "claim": "among the frameworks added, one built from 499 publicly reported incidents", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the unit and its evidence: a risk as a framework states it", "date": "2025-12-04 (post date)", "quote_ok": true, "checked_against": "saved text B7.txt (sha256 21849cbdbd3f0259…), whitespace-normalised" }, "F34": { "source": "B8", "origin": "card", "locus": "Who we are ¶1", "quote": "The MIT AI Risk Initiative (AIRI) provides authoritative data and frameworks to help decision-makers across the AI ecosystem identify, prioritize, and manage risks from AI.", "claim": "the operator: the MIT AI Risk Initiative (AIRI)", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "operator: the MIT AI Risk Initiative at MIT FutureTech", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 a7b938e0aa7923a9…), whitespace-normalised" }, "F35": { "source": "B8", "origin": "card", "locus": "Who we are ¶2", "quote": "We are based at MIT FutureTech, a joint initiative of the Institute of the Digital Economy and MIT CSAIL at the Massachusetts Institute of Technology", "claim": "based at MIT FutureTech, a joint initiative of the IDE and MIT CSAIL", "modality": "documented", "kind": "attribution", "kernel": true, "const": false, "lineage": "operator: the MIT AI Risk Initiative at MIT FutureTech", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 a7b938e0aa7923a9…), whitespace-normalised" }, "F36": { "source": "B8", "origin": "card", "locus": "What we do ¶1; AI Incident Tracker", "quote": "No single dataset captures all of this, so we maintain several connected workstreams, each answering a different question. … A catalogue that classifies incidents where AI has already caused harm. Developed in collaboration with the AI Incident Database.", "claim": "the repository one of several connected workstreams; the Incident Tracker another, developed with the AI Incident Database", "modality": "documented", "kind": "attribution", "kernel": true, "const": false, "lineage": "the Initiative's workstreams: the tracker beside the repository", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 a7b938e0aa7923a9…), whitespace-normalised" }, "F37": { "source": "B8", "origin": "card", "locus": "What we do: AI Risk Repository", "quote": "A living database of more than 1,600 risks from more than 65 frameworks and papers, taxonomies to navigate risks, and expert-led prioritization of the most pressing risks.", "claim": "the About page's count: more than 1,600 risks from more than 65 frameworks and papers", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 a7b938e0aa7923a9…), whitespace-normalised" }, "F38": { "source": "B8", "origin": "card", "locus": "Our mission ¶3", "quote": "Our flagship output, the AI Risk Repository, synthesizes 1700+ risks from 70+ frameworks into a common taxonomy.", "claim": "the same page's second count: 1700+ risks from 70+ frameworks, the flagship output", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "count and version: 777/43 to 1,700+/74", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 a7b938e0aa7923a9…), whitespace-normalised" }, "F39": { "source": "B8", "origin": "card", "locus": "Influence / Reach", "quote": "250+ … Citations … Including the 2025 and 2026 International AI Safety Reports … 290,000+ … Website users … One of the top sources returned for search queries on AI risks and incidents", "claim": "its reach: 250+ citations, 290,000+ website users, one of the top sources returned for search queries on AI risks and incidents", "modality": "self-description", "kind": "fact", "kernel": false, "const": false, "lineage": "reach: users, citations, search", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 a7b938e0aa7923a9…), whitespace-normalised" }, "F40": { "source": "B8", "origin": "card", "locus": "Who we are ¶3", "quote": "Our work is free to use and openly available, and most of our outputs are published under a CC BY 4.0 licence.", "claim": "free to use; most outputs CC BY 4.0", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "openness: free to copy, CC BY 4.0", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B8.txt (sha256 a7b938e0aa7923a9…), whitespace-normalised" }, "F41": { "source": "B9", "origin": "card", "locus": "post (the card's snippet)", "quote": "The initiative includes the MIT AI Risk Repository, a living database of more than 1,700 AI risks, the AI Incident Tracker, a collaboration with the Responsible AI Collaborative, which connects risks to over 1,400 incidents", "claim": "the Initiative's post: the repository (1,700+ risks) and the tracker, a collaboration with the Responsible AI Collaborative, over 1,400 incidents", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "the Initiative's workstreams: the tracker beside the repository", "date": "2026-10-29 (post undated as fetched)", "quote_ok": true, "checked_against": "saved text B9.txt (sha256 ddf88ccb5ab1762a…), whitespace-normalised" }, "F42": { "source": "B9", "origin": "card", "locus": "page about", "quote": "The MIT AI Risk Repository is a comprehensive living database of over 1000 AI risks categorized by their cause and risk domain", "claim": "the LinkedIn page's description: over 1000 risks by cause and domain", "modality": "documented", "kind": "definition", "kernel": false, "const": false, "lineage": "what it is: a living database of AI risks from frameworks", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B9.txt (sha256 ddf88ccb5ab1762a…), whitespace-normalised" }, "F43": { "source": "B10", "origin": "card", "locus": "card snippet (ScienceDirect returned 403)", "quote": "Practical implications for technology management. Our AI Risk Repository provides concrete tools for managing AI's societal impact across multiple stakeholder", "claim": "the published article's snippet: concrete tools for managing AI's societal impact across stakeholders", "modality": "attributed", "kind": "function", "kernel": false, "const": false, "lineage": "purpose: a common frame of reference for AI risk", "date": "2026 (Patterns, by B6)", "quote_ok": true, "checked_against": "the card's title/snippet in the registry's cite_list (saved as B10.txt)" }, "F44": { "source": "B11", "origin": "own_page", "locus": "What is the AI Risk Repository?", "quote": "The AI Risk Repository has three parts: … The AI Risk Database captures 1700+ risks extracted from 74 existing frameworks and classifications of AI risks", "claim": "the repository's own page: three parts; the database captures 1700+ risks from 74 frameworks", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "what it is: a living database of AI risks from frameworks", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F45": { "source": "B11", "origin": "own_page", "locus": "Domain Taxonomy of AI Risks", "quote": "The Domain Taxonomy of AI Risks classifies risks from AI into seven domains and 24 subdomains.", "claim": "seven domains and 24 subdomains, 7.6 multi-agent risks the last", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the Domain Taxonomy: seven domains and their subdomains", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F46": { "source": "B11", "origin": "own_page", "locus": "Causal Taxonomy of AI Risks", "quote": "AI: Due to a decision or action made by an AI system … Human: Due to a decision or action made by humans … Other: Due to some other reason or ambiguous", "claim": "each causal factor has a third value, Other (entity shown; intent and timing the same)", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the Causal Taxonomy: entity, intent, timing", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F47": { "source": "B11", "origin": "own_page", "locus": "FAQ: How did you create the AI Risk Repository?", "quote": "We used a systematic search strategy, forwards and backwards searching, and expert consultation, to identify 65 AI risk classifications, frameworks, and taxonomies. We extracted 1600+ risks from these documents into a living AI risk database.", "claim": "method: systematic search, forwards and backwards searching, expert consultation; 65 documents, 1600+ risks", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "method: systematic review, expert consultation, best-fit synthesis", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F48": { "source": "B11", "origin": "own_page", "locus": "FAQ: limitations", "quote": "Has potential for errors and subjective bias; we used a single expert reviewer for extraction and coding. … May include poorly communicated or unclear risks: we extracted risks as presented.", "claim": "limits: a single expert reviewer for extraction and coding; risks extracted as presented", "modality": "self-description", "kind": "limit", "kernel": true, "const": false, "lineage": "stated limits", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F49": { "source": "B11", "origin": "own_page", "locus": "FAQ: limitations", "quote": "Limited to risks from 65 documents (although we screened >17,000 records after a systematic search of peer-reviewed and gray literature) … May be missing emerging, domain-specific risks, and unpublished risks.", "claim": "limits: 65 documents (from >17,000 screened); may miss emerging, domain-specific and unpublished risks", "modality": "self-description", "kind": "limit", "kernel": false, "const": false, "lineage": "stated limits", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F50": { "source": "B11", "origin": "own_page", "locus": "FAQ: limitations", "quote": "Our taxonomies prioritize clarity and simplicity over nuance. … Our taxonomies do not categorize risks by potentially important factors such as risk impact, likelihood, or discuss the interaction between risks.", "claim": "limits: clarity over nuance; no impact, likelihood or interaction", "modality": "self-description", "kind": "limit", "kernel": false, "const": false, "lineage": "stated limits", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F51": { "source": "B11", "origin": "own_page", "locus": "How to use the AI Risk Repository", "quote": "Our Database is free to copy and use", "claim": "the database free to copy and use", "modality": "stipulation", "kind": "fact", "kernel": false, "const": false, "lineage": "openness: free to copy, CC BY 4.0", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F52": { "source": "B11", "origin": "own_page", "locus": "FAQ: other databases", "quote": "What are some other databases of AI risks? … https://attack.mitre.org/ … https://airisk.io/ … https://avidml.org/#efforts … https://www.aitracker.org/#catalog-tabs", "claim": "other databases of AI risks named by the repository: MITRE ATT&CK, airisk.io, AVID, aitracker.org", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "place: other databases of AI risks, the harm basin", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F53": { "source": "B11", "origin": "own_page", "locus": "FAQ: two taxonomies", "quote": "Because these classification systems were so different, it was hard to unify them; high-level risk categories such as “Diffusion of responsibility” or “Humans create dangerous AI by mistake” do not map to narrower categories like “Misuse” or “Noisy Training Data,” or vice versa.", "claim": "why two taxonomies: causal categories and hazard categories do not map onto each other", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "method: systematic review, expert consultation, best-fit synthesis", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F54": { "source": "B11", "origin": "own_page", "locus": "FAQ: citation", "quote": "To reference our repository, you can cite our pre-print paper:", "claim": "the repository cited by its preprint", "modality": "stipulation", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the paper: Slattery et al., arXiv:2408.12622, Patterns 2026", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B11.txt (sha256 2b705e1a289760fe…), whitespace-normalised" }, "F55": { "source": "B12", "origin": "own_page", "locus": "¶2", "quote": "Use this file to orient LLMs and AI agents to the main AIRI resources at airisk.mit.edu and the related AI Risk Navigator.", "claim": "a file addressed to LLMs and AI agents, orienting them to the Initiative's resources", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "machine-facing guidance: llms.txt", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B12.txt (sha256 039e65007ed87700…), whitespace-normalised" }, "F56": { "source": "B12", "origin": "own_page", "locus": "Important guidance for AI assistants", "quote": "Refer to the project as the MIT AI Risk Initiative, or AIRI when the abbreviation is useful. … Treat AIRI resources as living datasets. Check the linked pages before giving exact counts, dates, version numbers, or coverage claims.", "claim": "guidance to assistants: name the MIT AI Risk Initiative; check the pages before giving exact counts, dates or versions", "modality": "stipulation", "kind": "function", "kernel": true, "const": false, "lineage": "machine-facing guidance: llms.txt", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B12.txt (sha256 039e65007ed87700…), whitespace-normalised" }, "F57": { "source": "B12", "origin": "own_page", "locus": "Important guidance for AI assistants", "quote": "Do not present AIRI datasets as legal, safety, or investment advice. They are research resources and decision-support tools.", "claim": "guidance: not legal, safety or investment advice; research resources", "modality": "stipulation", "kind": "limit", "kernel": false, "const": false, "lineage": "machine-facing guidance: llms.txt", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B12.txt (sha256 039e65007ed87700…), whitespace-normalised" }, "F58": { "source": "B12", "origin": "own_page", "locus": "Suggested Answering Behavior", "quote": "For questions about observed harms, use the MIT AI Incident Tracker and distinguish reported incidents from predicted risks. … For cross-dataset questions, use the MIT AI Risk Navigator, but note that it is a research prototype and may contain preliminary analyses.", "claim": "guidance: distinguish reported incidents from predicted risks; note that the Navigator is a research prototype", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "machine-facing guidance: llms.txt", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B12.txt (sha256 039e65007ed87700…), whitespace-normalised" }, "F59": { "source": "B12", "origin": "own_page", "locus": "Core AIRI Pages", "quote": "[AI Risk Repository](https://airisk.mit.edu/risks): Living database of AI risks, including the AI Risk Database, Causal Taxonomy, Domain Taxonomy, methodology notes, use cases, limitations, data access links, and citation guidance.", "claim": "the canonical page for the repository is /risks, with its limitations and citation guidance", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "what it is: a living database of AI risks from frameworks", "date": "2026-10-29 (page undated; as fetched)", "quote_ok": true, "checked_against": "saved text B12.txt (sha256 039e65007ed87700…), whitespace-normalised" } }, "field_sources": { "B1": "MIT AI Risk Initiative — MIT AI Risk Initiative", "B2": "MIT AI Risk Initiative — AI Risk Repository Report updated (April 2025)", "B3": "MIT Initiative on the Digital Economy — MIT Researchers Create an AI Risk Repository - MIT Initiative on the Digital Economy", "B4": "MIT FutureTech — FutureTech", "B5": "MIT AI Risk Navigator — MIT AI Risk Navigator", "B6": "arXiv — [2408.12622] The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence", "B7": "MIT AI Risk Initiative — Repository Update: December 2025", "B8": "MIT AI Risk Initiative — About", "B9": "LinkedIn — MIT AI Risk Initiative | LinkedIn", "B10": "ScienceDirect.com — The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence - ScienceDirect", "B11": "MIT AI Risk Initiative — AI Risk Repository (airisk.mit.edu/risks)", "B12": "MIT AI Risk Initiative — llms.txt (airisk.mit.edu/llms.txt)", "W1": "Wikipedia — no article" }, "delta": { "T_vs_LB": "At 'mit ai risk repository' (T1, ten cards: B1–B10) the composition gives the repository by MIT's pages and the paper's abstract. Composed from its cards at their grain: the living database (B1 F2; B9 F42), 'developed by MIT FutureTech' (B3 F11 'A new MIT FutureTech project'; B8 F35), the causal taxonomy by entity, intent and timing (B4 F19), the seven domains named as the v1 abstract names them with the repository page's 24 subdomains (B4 F20; B11 F45, an own page not surfaced), the Navigator as a connected tool (B5 F22) and the tracker's collaboration with the AI Incident Database (B8 F36). Substituted reading: T1-4 composes 'roughly 38% to 42% of risks stem directly from AI systems' where the paper (B6 F28) gives 38% to human decisions and 42% to AI systems: the human share is folded into the AI range, and 'contrary to common assumptions' (the finding's point, that humans cause nearly as many as AI) is not composed; the first release's 51% and 34% (B3 F14) are not composed either. The finding is composed at the source's resolution (its figures) with no source named in the sentence, markers only (§6.2: at the source's own resolution, recorded as the grain rule's case). One count where the field holds several: T1-2's 'over 1,700 … more than 65 published frameworks' is the home page's figure (B1 F3; B8 F37 'more than 65'), while the version 4 post adds 9 frameworks to version 3's 65 (B7 F32; B2 F6) and the arXiv card's page and the repository's own page give 74 (B6 F27; B11 F44); the composition gives no version and no date for the count, against the entity's own guidance to assistants to 'Check the linked pages before giving exact counts, dates, version numbers, or coverage claims' (B12 F56, an own page not surfaced; the guidance audits and does not substitute, §3.5). Limit dropped: the causal taxonomy loses each factor's third value, Other (B11 F46); the Navigator loses 'This is a research prototype. The data and analyses are preliminary and not yet validated' (B5 F23; also B12 F58), and is composed as linking the repository to incidents, governance frameworks and mitigations, at the grain of its page (B5 F22; 'EU AI Act' appears on the page outside the ledgered spans). Composed from the card's snippet rather than its page: 'catalogs and standardizes' (T1-1) is the arXiv card's own snippet, 'Overview: The AI Risk Repository catalogs and standardizes fragmented artificial intelligence risk terminology', not the abstract's wording ('organized them into a unified system', F27); the IDE card's snippet is of the same kind ('Overview: MIT FutureTech’s AI Risk Repository catalogs 777 risks…'). 'open-access' is at the grain of the field's licence (B1 F5 CC BY 4.0; B11 F51 'free to copy and use'), the licence unnamed. Operator: T names MIT FutureTech and never the MIT AI Risk Initiative, which every MIT card's site line carries and the About page states as the operator (B8 F34); the entity's guidance asks assistants to 'Refer to the project as the MIT AI Risk Initiative' (B12 F56). The tracker's collaboration with the Responsible AI Collaborative (B9 F41) is not composed. Available, not composed: the unit and its evidence, each risk linked to source, quotes and page numbers (B3 F12; B1 F4); the method (B4 F18; B11 F47, F53); the purpose, shared terminology (B6 F25–F26) and common frame (B1 F1; B4 F21; B10 F43); the gap finding, 34% of subdomains per framework (B3 F13); every stated limit, a single expert reviewer, risks extracted as presented, 65 documents, no impact or likelihood (B11 F48–F50; B3 F15); the version history (B2 F6–F8, F10; B3 F11; B4 F17; B7 F31–F33); the paper's authors, its versions and its publication in Patterns (B6 F29–F30; B11 F54); the About page's count of 1700+ from 70+ beside 1,600 from 65+ (F37–F38); the reach and the About page's 'One of the top sources returned for search queries on AI risks and incidents' (B8 F39); the Initiative's other workstreams (B8 F36). Content replaced by use: the closer (T1-8) offers the repository for 'a specific industry audit, regulatory compliance check (such as the EU AI Act), or risk assessment'; the field's use cases include audits and risk assessments (B11, outside the ledgered spans), and the entity's guidance says 'Do not present AIRI datasets as legal, safety, or investment advice. They are research resources and decision-support tools.' (B12 F57): the closer and the guidance are recorded side by side, no breach asserted. No card cites the MMRS record (the registry's q_kind: 'no card cites it'); no card names the archive.", "LB_vs_LBA": "Admission adds, in one claim space with the field's: the unit read as 'the risk as a framework states it' (MMRS-03, MMRS-12), on the field's own object (F12, F4, MMRS-04 one lineage with F4); what the selection gathers and leaves out, the archive's sense the field states only as a limit: 'It aggregates aggregators', 'What enters is what a framework named', 'A risk no framework has written down has no row' (MMRS-06–08), beside the FAQ's 'May be missing emerging, domain-specific risks, and unpublished risks' (F49) and a framework added from 499 incidents (F33); the repository placed over the harm basin as a taxonomy for it (MMRS-10), its role harm and its neighbour the AI Incident Database (MMRS-18, MMRS-19, MMRS-22); the incident layer read as a second-order machine reading of AIID (MMRS-09, MMRS-21) and the incidents kept apart from the risks (MMRS-14); the record kept apart from the entity, each assertion with a basis, lifecycle apart from observation (MMRS-25), with its source and its date (MMRS-16) and its unchecked items (MMRS-17), which the field answers (method F18, F47; licence F5, F40, F51, F30); the Navigator held as a distinct entity and a census candidate (MMRS-23, MMRS-24, CAP-1-05, CAP-1-06); and the composition layer's observation at the address (CAP-1-01–04, MMRS-26). Coherence (P_coh, composed each with its source): the record's 65 frameworks (MMRS-01, MMRS-13) and the Capture Registry's reading of 1,700 from 65 as version 4's and 'the current one' (CAP-1-02, CAP-1-04) beside 74 in the paper's v3 and on the repository's page (F27, F44) and the 9 frameworks version 4 added to version 3's 65 (F32, F6); the record's tracker as part of the repository (MMRS-20, MMRS-02 'its AI Incident Tracker', MMRS-14) beside the field's tracker as a sibling workstream of the Initiative (F36, F41). Nothing the archive admits denies a field presupposition: the field's object (a database of risks from frameworks, with two taxonomies, run by the Initiative) is the archive's object, and the archive's additions are senses, limits and relations on it. Not surfaced: the MMRS record and the Capture Registry (no card); the repository's own page (/risks) and llms.txt, own pages, were not surfaced by T either and are admitted by the operator's ruling." }, "kernel": [ { "K": "K1", "claim": "`MMRS-01`", "source": "MMRS record", "M_src": "stipulation", "sense": "what it is: a living database of AI risks from frameworks", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "rival claim (F27, F44: 74 frameworks for version 4)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K2", "claim": "`MMRS-02`", "source": "MMRS record", "M_src": "documented", "sense": "the Initiative's workstreams: the tracker beside the repository", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F36, F41)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K3", "claim": "`MMRS-03`", "source": "MMRS record", "M_src": "stipulation", "sense": "the unit and its evidence: a risk as a framework states it", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K4", "claim": "`MMRS-04`", "source": "MMRS record", "M_src": "documented", "sense": "the unit and its evidence: a risk as a framework states it", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (one lineage with F4)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K5", "claim": "`MMRS-05`", "source": "MMRS record", "M_src": "documented", "sense": "the Domain Taxonomy: seven domains and their subdomains", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (one lineage with F7)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K6", "claim": "`MMRS-06`", "source": "MMRS record", "M_src": "interpretation", "sense": "what it gathers: aggregates aggregators", "qualifiers carried": "F48, F49, F50 (the field's stated limits)", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K7", "claim": "`MMRS-07`", "source": "MMRS record", "M_src": "interpretation", "sense": "what it gathers: aggregates aggregators", "qualifiers carried": "F48, F49, F50 (the field's stated limits)", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F49)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K8", "claim": "`MMRS-08`", "source": "MMRS record", "M_src": "interpretation", "sense": "what it gathers: aggregates aggregators", "qualifiers carried": "F48, F49, F50 (the field's stated limits)", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F49: may be missing unpublished risks)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K9", "claim": "`MMRS-09`", "source": "MMRS record", "M_src": "interpretation", "sense": "the Initiative's workstreams: the tracker beside the repository", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K10", "claim": "`MMRS-10`", "source": "MMRS record", "M_src": "interpretation", "sense": "place: other databases of AI risks, the harm basin", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K11", "claim": "`MMRS-11`", "source": "MMRS record", "M_src": "attributed", "sense": "operator: the MIT AI Risk Initiative at MIT FutureTech", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (one lineage with F34, F35)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K12", "claim": "`MMRS-12`", "source": "MMRS record", "M_src": "attributed", "sense": "the unit and its evidence: a risk as a framework states it", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (one lineage with F12)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K13", "claim": "`MMRS-13`", "source": "MMRS record", "M_src": "attributed", "sense": "count and version: 777/43 to 1,700+/74", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "rival claim (F27, F44)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K14", "claim": "`MMRS-14`", "source": "MMRS record", "M_src": "interpretation", "sense": "the Initiative's workstreams: the tracker beside the repository", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F36)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K15", "claim": "`MMRS-15`", "source": "MMRS record", "M_src": "documented", "sense": "count and version: 777/43 to 1,700+/74", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (one lineage with F31)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K16", "claim": "`MMRS-16`", "source": "MMRS record", "M_src": "documented", "sense": "the record kept apart from the entity", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K17", "claim": "`MMRS-17`", "source": "MMRS record", "M_src": "self-description", "sense": "the record kept apart from the entity", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F5, F18, F47: the unchecked items stated in the field)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K18", "claim": "`MMRS-18`", "source": "MMRS record", "M_src": "stipulation", "sense": "place: other databases of AI risks, the harm basin", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K19", "claim": "`MMRS-19`", "source": "MMRS record", "M_src": "stipulation", "sense": "place: other databases of AI risks, the harm basin", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K20", "claim": "`MMRS-20`", "source": "MMRS record", "M_src": "stipulation", "sense": "the Initiative's workstreams: the tracker beside the repository", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "rival claim (F36, F41: the tracker a workstream of the Initiative)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K21", "claim": "`MMRS-21`", "source": "MMRS record", "M_src": "attributed", "sense": "the Initiative's workstreams: the tracker beside the repository", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K22", "claim": "`MMRS-22`", "source": "MMRS record", "M_src": "interpretation", "sense": "place: other databases of AI risks, the harm basin", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K23", "claim": "`MMRS-23`", "source": "MMRS record", "M_src": "documented", "sense": "neighbour: the MIT AI Risk Navigator (research prototype)", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the Navigator a distinct entity)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K24", "claim": "`MMRS-24`", "source": "MMRS record", "M_src": "stipulation", "sense": "neighbour: the MIT AI Risk Navigator (research prototype)", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K25", "claim": "`MMRS-25`", "source": "MMRS record", "M_src": "stipulation", "sense": "the record kept apart from the entity", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K26", "claim": "`MMRS-26`", "source": "MMRS record", "M_src": "documented", "sense": "reception: the composition layer", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K27", "claim": "`CAP-1-01`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "reception: the composition layer", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K28", "claim": "`CAP-1-02`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "reception: the composition layer", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "rival claim (F27, F44)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K29", "claim": "`CAP-1-03`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "reception: the composition layer", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K30", "claim": "`CAP-1-04`", "source": "Capture Registry, 2026-10-29", "M_src": "interpretation", "sense": "count and version: 777/43 to 1,700+/74", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "rival claim (F27, F44)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K31", "claim": "`CAP-1-05`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "neighbour: the MIT AI Risk Navigator (research prototype)", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the Navigator a distinct entity)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K32", "claim": "`CAP-1-06`", "source": "Capture Registry, 2026-10-29", "M_src": "documented", "sense": "neighbour: the MIT AI Risk Navigator (research prototype)", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing finding (an observation of the composition at the entity's address)", "resolver": "a later capture at 'mit ai risk repository' (ADDR-5762c6084031) on Google AI Overview, seated in the Capture Registry (§8.3)" } ], "kernel_T": [ { "claim": "F28", "note": "the paper's finding, human decisions 38% and AI systems 42% (T1-4 substituted: '38% to 42% … from AI systems')" }, { "claim": "F27", "note": "74 frameworks in v3 (T1-2 composed 65)" }, { "claim": "F44", "note": "the repository's own page: 1700+ from 74" }, { "claim": "F46", "note": "each causal factor's third value, Other" }, { "claim": "F23", "note": "the Navigator a research prototype, not yet validated" }, { "claim": "F48", "note": "a single expert reviewer for extraction and coding" }, { "claim": "F12", "note": "each risk linked to its source and evidence" }, { "claim": "F34", "note": "the operator, the MIT AI Risk Initiative, unnamed in T" }, { "claim": "F56", "note": "the entity's guidance to check counts, dates and versions" }, { "claim": "F57", "note": "the entity's guidance: not legal or safety advice (beside T1-8)" }, { "claim": "F41", "note": "the tracker with the Responsible AI Collaborative" } ], "ledger": { "archive": "datasets/negative-of-the-negative/v2/ledgers/mit-ai-risk-repository/ledger-archive.json", "field": "datasets/negative-of-the-negative/v2/ledgers/mit-ai-risk-repository/field.json", "selection": "datasets/negative-of-the-negative/v2/traversal/mit-ai-risk-repository/", "config": "datasets/negative-of-the-negative/v2/panel/configs/mit-ai-risk-repository.json", "intake": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/", "audit": null }, "ontology": { "composition": "equal", "composition_note": "one claim space; every field claim composed as received beside the archive claims it stands with; nothing translated", "case": { "value": "equal", "statement": "The MMRS record holds the field's entity as the entity: it describes the repository from its own pages with each field's basis, and its article composes nothing that denies a field presupposition. What it adds is on the field's object: the unit read as the risk as a framework states it, what the selection gathers and so leaves out (a risk no framework has written down has no row), the repository as a taxonomy over the harm basin, the tracker as a second-order machine reading of the AI Incident Database, the record kept apart from the entity with its unchecked items, and the Navigator held apart as a candidate; the Capture Registry adds the composition layer's observation at the name. The ontology the archive uses is available in the field: the repository states as its own limit that it 'May be missing emerging, domain-specific risks, and unpublished risks' (F49) and that it extracted 'risks as presented' (F48), and the archive's 'What enters is what a framework named' extends that limit into a property of the unit. The law of the suppressed witness does not apply (§0.10: not where the field holds the ontology). Not mixed.", "loci": [ { "dep": "MMRS", "locus": "article, kernel s1 (articles.py)", "quote": "The MIT AI Risk Repository is a database of more than 1,700 AI risks extracted from 65 published risk frameworks, run by the MIT AI Risk Initiative (MIT FutureTech)." }, { "dep": "MMRS", "locus": "article, What it gathers s1–s3 (articles.py)", "quote": "It aggregates aggregators: its sources are frameworks, which are themselves syntheses. What enters is what a framework named. A risk no framework has written down has no row." }, { "dep": "MMRS", "locus": "census.json, schema_note", "quote": "each page has two nodes, the MMRS record and the entity it describes; relations are typed; each assertion carries a source and a basis (operator statement, third-party report, our observation, our reading); lifecycle is kept apart from the observation that saw it" }, { "dep": "capture", "locus": "mit-ai-risk-repository-aio-20261029, analysis", "quote": "The repository is composed at a finer grain than the record holds, its taxonomy named domain by domain, and the composition surfaces a sibling the census lacked." }, { "dep": "B11", "locus": "FAQ: limitations (field)", "quote": "May be missing emerging, domain-specific risks, and unpublished risks." } ] }, "E_F": { "type": "C (public entity: a database)", "definition": "a living database of AI risks extracted from published AI risk frameworks (version 4, December 2025: over 1,700 coded risks; 74 frameworks in the paper's v3 and on its own page, 65 on the home page), each risk linked to its source and evidence, classified by a Causal Taxonomy (entity, intentionality, timing) and a Domain Taxonomy (seven domains, 24 subdomains), built by systematic review, expert consultation and best-fit framework synthesis, the flagship of the MIT AI Risk Initiative at MIT FutureTech, described in Slattery et al., arXiv:2408.12622 (2024; v3 2026; Patterns 2026), CC BY 4.0, with stated limits (one expert reviewer; risks as presented); no Wikipedia article (B1–B12, W1)", "claims": [ "F2", "F44", "F32", "F27", "F12", "F19", "F20", "F45", "F18", "F34", "F35", "F29", "F30", "F5", "F48" ] }, "E_A": { "definition": "the same database, in the MMRS meta-aggregator's record: a database of more than 1,700 AI risks from published risk frameworks, run by the MIT AI Risk Initiative (MIT FutureTech), whose unit is the risk as a framework states it, an aggregator of aggregators in which a risk no framework has written down has no row, sitting over the harm basin as a taxonomy for it, with an incident layer that reads the AI Incident Database by machine; a record kept apart from the entity, its extraction method and licence unchecked; observed by the Capture Registry at its name, where the composition layer composed it at a finer grain than the record and surfaced the Navigator, a sibling the census lacked", "claims": [ "MMRS-01", "MMRS-03", "MMRS-06", "MMRS-08", "MMRS-10", "MMRS-09", "MMRS-25", "MMRS-17", "CAP-1-01", "CAP-1-05" ], "defining_papers": [], "defining_records": [ "MMRS record (census.json rows[slug == 'mit-ai-risk-repository'], candidates, schema_note; rows[slug == 'mit-ai-incident-tracker'].relations and intake; articles.py ARTICLES['mit-ai-risk-repository'] and ARTICLES['ai-incident-database'] Place in the field; rendered aggregators/mit-ai-risk-repository/index.html)", "Capture Registry mit-ai-risk-repository-aio-20261029 (OBS-21a7cea64b70)" ] }, "rho": { "value": "extends", "status": "read from the MMRS record and the capture record (§1.2a); no deposit names the entity or its operator (D/R/O string pass over 1,711 texts: zero candidates)", "statement": "the archive reads E_F as the entity and adds senses, limits and relations on the field's own object: the unit as the risk a framework states, what the selection gathers and leaves out, the taxonomy over the harm basin, the incident layer as a machine reading of AIID, the record kept apart from the entity with its unchecked items, the Navigator held apart, and the composition layer's observed treatment of the entity", "claims": [ "MMRS-03", "MMRS-06", "MMRS-07", "MMRS-08", "MMRS-10", "MMRS-09", "MMRS-25", "MMRS-17", "MMRS-23", "CAP-1-01" ], "guard": "ρ = extends claims an addition and no more: MMRS-06–10 are the record's readings (interpretation), not findings that the repository is defective; where an archive claim meets a field claim (MMRS-01 and MMRS-13 and CAP-1-02/04 beside F27 and F44; MMRS-20 beside F36 and F41) both are composed under P_coh in the one claim space", "considered": [ { "value": "constructs", "set_aside": "the archive does not read the repository as made by its receiving texts" }, { "value": "revises", "set_aside": "the entity is not the archive's own" }, { "value": "witness (suppressed witness)", "set_aside": "the field holds the ontology (§0.10)" } ] }, "translation": [ { "lineage": "purpose: a common frame of reference for AI risk", "field_claims": [ "F1", "F21", "F25", "F26", "F43" ], "received": "F1 (B1, hero ¶1–2): \"Producing the information needed to address risks from AI … Authoritative data and frameworks to help you identify, prioritize, and manage the risks from AI\"; F21 (B4, abstract): \"The AI Risk Repository is, to our knowledge, the first attempt to rigorously curate, analyze, and extract AI risk frameworks into a publicly accessible, comprehensive, extensible, and categorized risk database.\"; F25 (B6, Abstract): \"Yet researchers, policymakers, and technology companies lack shared terminology for discussing AI risks.\"; F26 (B6, Abstract): \"Consider \"privacy\": one framework uses this term to describe a model's ability to leak sensitive training data, while another uses it to mean freedom from government surveillance.\"; F43 (B10, card snippet (ScienceDirect returned 403)): \"Practical implications for technology management. Our AI Risk Repository provides concrete tools for managing AI's societal impact across multiple stakeholder\"", "receiving_text": "—", "translation": "untranslated: composed as received (no archive claim on this lineage)", "archive_claims": [] }, { "lineage": "what it is: a living database of AI risks from frameworks", "field_claims": [ "F2", "F9", "F42", "F44", "F59" ], "received": "F2 (B1, Our tools and data: 'What are the risks from AI?'): \"Learn about the many different AI risks and how they emerge with our living repository of AI risks\"; F9 (B2, About the AI Risk Repository ¶1): \"The AI Risk Repository is a structured and evolving resource that aims to consolidate and clarify how risks from artificial intelligence are categorized across academic, industry, and policy literatures.\"; F42 (B9, page about): \"The MIT AI Risk Repository is a comprehensive living database of over 1000 AI risks categorized by their cause and risk domain\"; F44 (B11, What is the AI Risk Repository?): \"The AI Risk Repository has three parts: … The AI Risk Database captures 1700+ risks extracted from 74 existing frameworks and classifications of AI risks\"; F59 (B12, Core AIRI Pages): \"[AI Risk Repository](https://airisk.mit.edu/risks): Living database of AI risks, including the AI Risk Database, Causal Taxonomy, Domain Taxonomy, methodology notes, use cases, limitations, data access links, and citation guidance.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-01", "archive_claims": [ "MMRS-01" ] }, { "lineage": "count and version: 777/43 to 1,700+/74", "field_claims": [ "F3", "F6", "F11", "F17", "F27", "F31", "F32", "F37", "F38" ], "received": "F3 (B1, All collections: AI Risks): \"A living database of 1,700+ AI risks from 65 frameworks\"; F6 (B2, What's new in this update): \"Integration of 22 new AI risk frameworks, bringing the total number of included documents to 65. … Expansion of the AI Risk Database to 1,612 unique risk entries, each systematically extracted and coded.\"; F11 (B3, lede ¶1): \"A new MIT FutureTech project reviewed 43 AI frameworks produced by research, industry and government organizations; they identified 777 risks in total.\"; F17 (B4, abstract): \"This comprises a living database of 777 risks extracted from 43 taxonomies, which can be filtered based on two overarching taxonomies and easily accessed, modified, and updated via our website and online spreadsheets.\"; F27 (B6, Abstract): \"We systematically analyzed every major AI risk framework published to date-74 frameworks containing 1,725 distinct risks-and organized them into a unified system.\"; F31 (B7, ¶1): \"We’re pleased to share that Version 4 of the MIT AI Risk Repository is now live.\"; F32 (B7, What's new in Version 4): \"9 newly added frameworks, spanning government & industry reports, peer reviewed journal articles, conference papers and preprints, … ~200 new AI risk categories, expanding the repository to over 1,700 coded risks\"; F37 (B8, What we do: AI Risk Repository): \"A living database of more than 1,600 risks from more than 65 frameworks and papers, taxonomies to navigate risks, and expert-led prioritization of the most pressing risks.\"; F38 (B8, Our mission ¶3): \"Our flagship output, the AI Risk Repository, synthesizes 1700+ risks from 70+ frameworks into a common taxonomy.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-13, MMRS-15, MMRS-16, CAP-1-02, CAP-1-04", "archive_claims": [ "MMRS-13", "MMRS-15", "MMRS-16", "CAP-1-02", "CAP-1-04" ] }, { "lineage": "the unit and its evidence: a risk as a framework states it", "field_claims": [ "F4", "F12", "F33" ], "received": "F4 (B1, blog: 'Explore the Frameworks Behind the AI Risk Repository'): \"This post shares a public Google Slides deck containing all 65 frameworks included in Version 3 of the AI Risk Repository. It provides direct access to source documents, diagrams, and key excerpts to support transparency, research, and further analysis of AI risk classifications.\"; F12 (B3, lede ¶2): \"The repository includes a risk database linking each risk to the source information (paper title, authors) and supporting evidence, such as quotes and page numbers.\"; F33 (B7, Newly added frameworks: Li et al., 2025): \"This preprint presents a taxonomy of Generative AI Failures based on a systematic analysis of 499 publicly reported AI incidents.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-03, MMRS-04, MMRS-12", "archive_claims": [ "MMRS-03", "MMRS-04", "MMRS-12" ] }, { "lineage": "openness: free to copy, CC BY 4.0", "field_claims": [ "F5", "F40", "F51" ], "received": "F5 (B1, footer): \"Data from the MIT AI Risk Initiative is licensed under CC BY 4.0\"; F40 (B8, Who we are ¶3): \"Our work is free to use and openly available, and most of our outputs are published under a CC BY 4.0 licence.\"; F51 (B11, How to use the AI Risk Repository): \"Our Database is free to copy and use\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-17", "archive_claims": [ "MMRS-17" ] }, { "lineage": "the Domain Taxonomy: seven domains and their subdomains", "field_claims": [ "F7", "F8", "F20", "F45" ], "received": "F7 (B2, What's new in this update): \"Introduction of a new risk subdomain on multi-agent risks, reflecting recent developments in AI research on complex agent interactions.\"; F8 (B2, About the AI Risk Repository): \"A Domain Taxonomy, which classifies risks into seven thematic domains and 25 more granular subdomains.\"; F20 (B4, abstract): \"Our mid-level Domain Taxonomy of AI Risks classifies risks into seven AI risk domains: (1) Discrimination & toxicity, (2) Privacy & security, (3) Misinformation, (4) Malicious actors & misuse, (5) Human-computer interaction, (6) Socioeconomic & environmental, and (7) AI system safety, failures, & limitations. These are further divided into 23 subdomains.\"; F45 (B11, Domain Taxonomy of AI Risks): \"The Domain Taxonomy of AI Risks classifies risks from AI into seven domains and 24 subdomains.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-05, CAP-1-02", "archive_claims": [ "MMRS-05", "CAP-1-02" ] }, { "lineage": "the paper: Slattery et al., arXiv:2408.12622, Patterns 2026", "field_claims": [ "F10", "F29", "F30", "F54" ], "received": "F10 (B2, header): \"Updated Version of the AI Risk Repository Preprint Now Available … Original release: August 2024 … Updated version: April 2025\"; F29 (B6, header; Authors): \"[Submitted on 14 Aug 2024 (v1), last revised 5 May 2026 (this version, v3)] … Authors:Peter Slattery, Alexander K. Saeri, Emily A. C. Grundy, Jess Graham, Michael Noetel, Risto Uuk, James Dao, Soroush Pour, Stephen Casper, Neil Thompson\"; F30 (B6, Comments; Journal reference): \"This paper has been published in Patterns (Cell Press, 2026) under a CC BY 4.0 licence … Patterns 101517 (2026)\"; F54 (B11, FAQ: citation): \"To reference our repository, you can cite our pre-print paper:\"", "receiving_text": "—", "translation": "untranslated: composed as received (no archive claim on this lineage)", "archive_claims": [] }, { "lineage": "headline findings: coverage gaps and causes", "field_claims": [ "F13", "F14", "F28" ], "received": "F13 (B3, interview: key findings): \"The key finding from our analysis is that there are significant gaps in existing risk frameworks, with the average framework covering only 34% of the identified risk subdomains and even the most significant frameworks covering only 70%.\"; F14 (B3, interview: key findings): \"most of the risks (51%) were caused by AI systems rather than humans (34%), and were found after the AI model was trained and deployed (65%) rather than before (10%).\"; F28 (B6, Abstract): \"Our two classification systems reveal important patterns: contrary to common assumptions, human decisions cause nearly as many AI risks (38%) as the AI systems themselves (42%).\"", "receiving_text": "—", "translation": "untranslated: composed as received (no archive claim on this lineage)", "archive_claims": [] }, { "lineage": "stated limits", "field_claims": [ "F15", "F48", "F49", "F50" ], "received": "F15 (B3, interview: audience): \"We expect that what we have produced will need some modification before it is useful for most audiences, but we hope that it provides a solid foundation.\"; F48 (B11, FAQ: limitations): \"Has potential for errors and subjective bias; we used a single expert reviewer for extraction and coding. … May include poorly communicated or unclear risks: we extracted risks as presented.\"; F49 (B11, FAQ: limitations): \"Limited to risks from 65 documents (although we screened >17,000 records after a systematic search of peer-reviewed and gray literature) … May be missing emerging, domain-specific risks, and unpublished risks.\"; F50 (B11, FAQ: limitations): \"Our taxonomies prioritize clarity and simplicity over nuance. … Our taxonomies do not categorize risks by potentially important factors such as risk impact, likelihood, or discuss the interaction between risks.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-06, MMRS-07, MMRS-08", "archive_claims": [ "MMRS-06", "MMRS-07", "MMRS-08" ] }, { "lineage": "reach: users, citations, search", "field_claims": [ "F16", "F39" ], "received": "F16 (B3, interview: reception): \"In less than a week, over 35,000 people have used the website and over 6,000 have viewed our explainer video on YouTube.\"; F39 (B8, Influence / Reach): \"250+ … Citations … Including the 2025 and 2026 International AI Safety Reports … 290,000+ … Website users … One of the top sources returned for search queries on AI risks and incidents\"", "receiving_text": "—", "translation": "untranslated: composed as received (no archive claim on this lineage)", "archive_claims": [] }, { "lineage": "method: systematic review, expert consultation, best-fit synthesis", "field_claims": [ "F18", "F47", "F53" ], "received": "F18 (B4, abstract): \"We construct our Repository with a systematic review of taxonomies and other structured classifications of AI risk followed by an expert consultation. We develop our taxonomies of AI risk using a best-fit framework synthesis.\"; F47 (B11, FAQ: How did you create the AI Risk Repository?): \"We used a systematic search strategy, forwards and backwards searching, and expert consultation, to identify 65 AI risk classifications, frameworks, and taxonomies. We extracted 1600+ risks from these documents into a living AI risk database.\"; F53 (B11, FAQ: two taxonomies): \"Because these classification systems were so different, it was hard to unify them; high-level risk categories such as “Diffusion of responsibility” or “Humans create dangerous AI by mistake” do not map to narrower categories like “Misuse” or “Noisy Training Data,” or vice versa.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-17", "archive_claims": [ "MMRS-17" ] }, { "lineage": "the Causal Taxonomy: entity, intent, timing", "field_claims": [ "F19", "F46" ], "received": "F19 (B4, abstract): \"Our high-level Causal Taxonomy of AI Risks classifies each risk by its causal factors (1) Entity: Human, AI; (2) Intentionality: Intentional, Unintentional; and (3) Timing: Pre-deployment; Post-deployment.\"; F46 (B11, Causal Taxonomy of AI Risks): \"AI: Due to a decision or action made by an AI system … Human: Due to a decision or action made by humans … Other: Due to some other reason or ambiguous\"", "receiving_text": "—", "translation": "untranslated: composed as received beside CAP-1-02", "archive_claims": [ "CAP-1-02" ] }, { "lineage": "neighbour: the MIT AI Risk Navigator (research prototype)", "field_claims": [ "F22", "F23", "F24" ], "received": "F22 (B5, hero): \"The Navigator connects MIT's AI Risk Repository datasets through a shared taxonomy, so you can move between them and surface the patterns that matter.\"; F23 (B5, banner): \"This is a research prototype. The data and analyses are preliminary and not yet validated\"; F24 (B5, hero ¶2; footer): \"A project of the AI Risk Initiative at MIT FutureTech, with the generous support of the Cambridge Boston Alignment Initiative. … © 2026 Spencer Michaels\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-23, MMRS-24, CAP-1-05, CAP-1-06", "archive_claims": [ "MMRS-23", "MMRS-24", "CAP-1-05", "CAP-1-06" ] }, { "lineage": "operator: the MIT AI Risk Initiative at MIT FutureTech", "field_claims": [ "F34", "F35" ], "received": "F34 (B8, Who we are ¶1): \"The MIT AI Risk Initiative (AIRI) provides authoritative data and frameworks to help decision-makers across the AI ecosystem identify, prioritize, and manage risks from AI.\"; F35 (B8, Who we are ¶2): \"We are based at MIT FutureTech, a joint initiative of the Institute of the Digital Economy and MIT CSAIL at the Massachusetts Institute of Technology\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-11", "archive_claims": [ "MMRS-11" ] }, { "lineage": "the Initiative's workstreams: the tracker beside the repository", "field_claims": [ "F36", "F41" ], "received": "F36 (B8, What we do ¶1; AI Incident Tracker): \"No single dataset captures all of this, so we maintain several connected workstreams, each answering a different question. … A catalogue that classifies incidents where AI has already caused harm. Developed in collaboration with the AI Incident Database.\"; F41 (B9, post (the card's snippet)): \"The initiative includes the MIT AI Risk Repository, a living database of more than 1,700 AI risks, the AI Incident Tracker, a collaboration with the Responsible AI Collaborative, which connects risks to over 1,400 incidents\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-02, MMRS-09, MMRS-14, MMRS-20, MMRS-21", "archive_claims": [ "MMRS-02", "MMRS-09", "MMRS-14", "MMRS-20", "MMRS-21" ] }, { "lineage": "place: other databases of AI risks, the harm basin", "field_claims": [ "F52" ], "received": "F52 (B11, FAQ: other databases): \"What are some other databases of AI risks? … https://attack.mitre.org/ … https://airisk.io/ … https://avidml.org/#efforts … https://www.aitracker.org/#catalog-tabs\"", "receiving_text": "—", "translation": "untranslated: composed as received beside MMRS-10, MMRS-18, MMRS-19, MMRS-22", "archive_claims": [ "MMRS-10", "MMRS-18", "MMRS-19", "MMRS-22" ] }, { "lineage": "machine-facing guidance: llms.txt", "field_claims": [ "F55", "F56", "F57", "F58" ], "received": "F55 (B12, ¶2): \"Use this file to orient LLMs and AI agents to the main AIRI resources at airisk.mit.edu and the related AI Risk Navigator.\"; F56 (B12, Important guidance for AI assistants): \"Refer to the project as the MIT AI Risk Initiative, or AIRI when the abbreviation is useful. … Treat AIRI resources as living datasets. Check the linked pages before giving exact counts, dates, version numbers, or coverage claims.\"; F57 (B12, Important guidance for AI assistants): \"Do not present AIRI datasets as legal, safety, or investment advice. They are research resources and decision-support tools.\"; F58 (B12, Suggested Answering Behavior): \"For questions about observed harms, use the MIT AI Incident Tracker and distinguish reported incidents from predicted risks. … For cross-dataset questions, use the MIT AI Risk Navigator, but note that it is a research prototype and may contain preliminary analyses.\"", "receiving_text": "—", "translation": "untranslated: composed as received (no archive claim on this lineage)", "archive_claims": [] } ], "receiving_texts": "equal-terms case: every field claim composed as received, beside the archive claims it stands with" }, "measure": { "rule": "v0.9 §4.10: KO_B sentences surviving in the KO at similarity ≥ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence)", "v0.9": { "surviving": 0, "of": 12, "positions": [ null, null, null, null, null, null, null, null, null, null, null, null ] }, "opening_run": { "rule": "v0.9 §5.10 (equal-terms case): the field-only KO's sentences as the KO's opening run in their own order", "breach": false }, "blocks": { "rule": "v0.9 §4.1 (equal-terms and mixed cases): the KO does not open with the field as a block or close with the archive as one; sentences typed by the ledgers of the claims they cite (F field-only, A archive-only, M both)", "sequence": "MMFFFMFFMFFMMFMMFFMMMFFAA", "field_only_head": 0, "archive_only_tail": 2, "breach": false }, "max_similarity": [ 0.178, 0.129, 0.105, 0.412, 0.422, 0.34, 0.449, 0.13, 0.121, 0.162, 0.154, 0.105 ], "ko_sentence_1_max_similarity_to_KO_B": 0.219 }, "homonyms": { "names": [], "basis": "the MMRS record carries no disambiguation for this entity (census row without a 'disambiguation' field); no card at T1 is a homonym: every card is MIT's, the paper's, or the published article. Near names kept apart: airisk.io, another database of AI risks the repository's own FAQ names (F52); 'AIRI', the Initiative's abbreviation (F34, F56). The sibling entities are distinct entities, kept as relations, never senses of the name: the MIT AI Incident Tracker (its own /non row, composed in parallel; the MMRS record's 'part_of', MMRS-20), the MIT AI Risk Navigator (a census candidate, MMRS-23; B5), the AI Incident Database (the record's neighbour, MMRS-19), and the Initiative's other workstreams (AI Governance Map, Mitigation Map, priorities survey; F36).", "distinct_entities": [ "mit-ai-incident-tracker", "MIT AI Risk Navigator (census candidate)", "ai-incident-database", "MIT AI Risk Initiative (the operator)", "airisk.io" ], "substitutes": { "T": [] }, "register": { "address_1": { "slug": "mit-ai-risk-repository-aio-20261029", "obs_id": "OBS-21a7cea64b70" } } }, "notes": [ "One address, Google AI Overview (expanded), signed out, incognito, 2026-10-29: mit-ai-risk-repository-aio-20261029 (OBS-21a7cea64b70, ADDR-5762c6084031, AXN:0AAF.CAPTURE.🚀🕌🔒🏠✖️🌉), ten cards. One composition; no other transcript in the alignment window (§2.3).", "Wikipedia: no article. The raw wikitext for 'MIT AI Risk Repository', 'AI Risk Repository', 'AI risk repository', 'MIT AI Risk Initiative' and 'MIT FutureTech' answers HTTP 404, the rendered page 404 ('Wikipedia does not have an article with this exact name.'), and Special:Search for the quoted name finds one article that mentions it, 'Voiceverse NFT plagiarism scandal' ('The MIT AI Risk Repository catalogued the incident as an example of \"economic and cultural devaluation…'), recorded under W1 and not ledgered.", "Fetching: every reachable page saved verbatim with curl -sSL -m 30 -A 'Mozilla/5.0' on 2026-10-29 between 05:08 and 05:12 UTC (the files' mtimes are local times), raw and text with sha256 in field.json. Card URLs found by WebSearch on title and site, or by the links on airisk.mit.edu (B2, B5, B7, B11) and the arXiv page's DOI (B10). Unreachable: B10, ScienceDirect (HTTP 403; cell.com, the publisher's other host, also 403); its snippet stands. Card 2 was first fetched at the wrong post of the same month ('Repository Update: April 2025', 1 April 2025) and refetched at the post whose title and opening match the card (23 April 2025); the saved B2 is the matching post.", "Own pages: the three the task names, airisk.mit.edu/ (B1), its About page (B8) and the arXiv abstract page (B6), are all cards at T1 and are carried once, as cards. The MMRS record's one source, airisk.mit.edu/, now opens on the MIT AI Risk Initiative's home, not on the repository; the repository's own page, airisk.mit.edu/risks (B11), and the site's llms.txt (B12) were added as own pages, not surfaced by T and not cited by the MMRS record. llms.txt is a source with its own summary policy addressed to assistants; under §3.5 the policy audits and does not substitute, so it is extracted by the same rule and its guidance is used in the delta as the entity's own statement, never as a rule the composition was bound by.", "The archive's deposits do not name the entity: the D/R/O string pass over 1,711 texts (aliases: AI Risk Repository, AI Risk Initiative, AIRI, airisk.mit.edu, MIT FutureTech, 2408.12622, Slattery; class: AI risk frameworks, taxonomies and databases, sourced from B6, B4, B11) returned zero candidates at D, R and O, so there was no hop. The archive's bearing is carried by its own records: the MMRS record (MMRS-01–26, including the tracker's census row and the AI Incident Database's article where they state claims on this entity, and the candidate entry for the Navigator) and the capture (CAP-1-01–06).", "Defects found, reported and not fixed. (1) MMRS kernel and census count 'more than 1,700 risks from 65 frameworks' with lifecycle 'version 4, December 2025': version 4 is 74 frameworks in the paper's v3 abstract (F27) and on the repository's own page (F44), and the December 2025 post adds 9 frameworks to version 3's 65 (F32, F6); 65 is the home page's figure (F3), the record's source; the About page itself gives both 65+ and 70+ (F37, F38). (2) Capture Registry, findings[0] and findings[1]: '1,700 risks from 65 frameworks is the record's count (version 4, December 2025)' and 'the composition gives the current one' carry the same mixed figure as current; findings[1] attributes 74 only to 'the arXiv overview card', though the repository's own page gives 74. (3) Capture Registry 'd' and 'd_full' restate T's 'with 38% to 42% of risks attributed to AI systems' without noting that the paper (F28) gives 38% to human decisions and 42% to AI systems; the findings do not record the substitution. (4) MMRS: the tracker's census row has 'part_of: mit-ai-risk-repository' (rendered on the repository's page as 'Has part MIT AI Incident Tracker') and the kernel says 'its AI Incident Tracker', while the About page, the LinkedIn post and llms.txt place the tracker beside the repository as a workstream of the Initiative (F36, F41, B12): contested, composed each with its source. (5) MMRS 'unchecked: extraction method, licence': both are stated on the entity's own pages (method F18, F47, F48; licence F5, F40, F51), available to the record's next reading. (6) The MMRS record for this entity has no 'disambiguation', 'negative_tags' or 'battery' fields (the brief names them for the census row); nothing to carry. (7) The MMRS record's URL airisk.mit.edu/ now opens on the Initiative's home page; the repository's canonical page per its own llms.txt is airisk.mit.edu/risks. (8) Within the field, MIT's own pages disagree on the subdomain count (23 in v1, 24 on /risks, 25 in the April 2025 report post: F20, F45, F8) and on the framework count (65, 70+, 74); recorded as field facts, composed each with its source.", "Composition: the composer read the ledgers; the transcript was read for T's claims and the delta only, as the brief asks. Ledgers unaudited (§3.8: one extractor). KO 25 sentences, every field claim (59) and archive claim (32) cited; KO_B 12 sentences, none surviving in the KO at ≥ 0.95; KO sentence 1 is mixed (MMRS-01 with F44, F2)." ], "composition_v09": { "source": "this row (composed directly; no rebuild/negative-of-the-negative/v09/mit-ai-risk-repository.json)", "composer": "Claude (Opus 5.5), subagent, 2026-10-29", "defining_material_read": [ { "what": "MMRS record, census row", "file": "/home/claude/fleet/machinemediation-org/data/meta-aggregator/census.json (rows[slug == 'mit-ai-risk-repository'], rows[slug == 'mit-ai-incident-tracker'], candidates, schema_note)", "read": "whole" }, { "what": "MMRS record, article", "file": "/home/claude/fleet/machinemediation-org/scripts/meta_aggregator/articles.py (ARTICLES['mit-ai-risk-repository']; ARTICLES['ai-incident-database'] and ['mit-ai-incident-tracker'] where they name the entity)", "read": "whole" }, { "what": "MMRS record, rendered", "file": "/home/claude/fleet/machinemediation-org/aggregators/mit-ai-risk-repository/index.html", "read": "text, for agreement with the sources" }, { "what": "capture record", "file": "data/EA-WG-CAPTURES-01.json (mit-ai-risk-repository-aio-20261029)", "read": "transcript, transcript_raw, cite_list, d, reading, findings, analysis, q_kind" }, { "what": "the specification", "file": "data/texts/AXN-0700-text.md", "read": "§0–§11 whole; Appendix A in part" } ], "cuts": [], "notes": "equal case; KO 25 sentences, all 59 field and 32 archive claims cited; P body 299 words; KO_B 12 sentences, none surviving at ≥ 0.95" }, "field_by_origin": { "note": "the field B kept apart by origin (brief §2): cards (surfaced by T), own pages (the repository's /risks and the site's llms.txt, admitted per the operator's ruling of 2026-10-29, not surfaced by T; the three own pages the task names are cards), Wikipedia (no article)", "cards": [ "B1", "B2", "B3", "B4", "B5", "B6", "B7", "B8", "B9", "B10" ], "own_pages": [ "B11", "B12" ], "wikipedia": [ "W1" ], "unreachable": [ "B10" ], "title_only": [], "wikipedia_state": "no article: raw wikitext 404 for 'MIT AI Risk Repository', 'AI Risk Repository', 'AI risk repository', 'MIT AI Risk Initiative', 'MIT FutureTech'; rendered page 404; Special:Search for the quoted name: one article citing the repository's classification of an incident (Voiceverse NFT plagiarism scandal), not an article on it", "counts": { "cards": 10, "card_sources": 10, "own_pages": 2, "wikipedia": 0, "fetched": 11, "unreachable": 1 }, "sources": [ { "id": "B1", "origin": "card", "role": "Overview card 1 at T1 (mit-ai-risk-repository-aio-20261029); also an own page (the MMRS record's source / named by the task)", "card": "MIT AI Risk Initiative — MIT AI Risk Initiative", "url": "https://airisk.mit.edu/", "kind": "Overview card (own page: the Initiative's home, the MMRS record's source)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 1 } ], "files": { "B1.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B1.html", "sha256": "d3f35366280a66cad3739c9cecb5216a73be687644466f0fabfd204af408be21" }, "B1.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B1.txt", "sha256": "935d8b0170045fd4112c26b32af21b397d0d9e2dfa05f6e14c1a7b3156cae3ac" } }, "fetched": { "at": "2026-10-29T05:08:14Z", "http": 200, "url": "https://airisk.mit.edu/", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B1.html", "raw_sha256": "d3f35366280a66cad3739c9cecb5216a73be687644466f0fabfd204af408be21", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B1.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "MIT AI Risk Initiative", "card_snippet": "Tools and Data: Expert Survey (expert consensus on severe AI risks, vulnerabilities, and responsibilities), AI Risk Repository (living repository of AI risks), ...", "unreachable": false, "names_operator": "yes ('MIT AI Risk Initiative', 'MIT FutureTech' in the saved text)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "the home page of airisk.mit.edu, now the MIT AI Risk Initiative's home; the card's snippet ('Tools and Data: Expert Survey … AI Risk Repository (living repository of AI risks)') is a condensed rendering of its 'Our tools and data' block, not a verbatim span of the saved text; the repository's own page is /risks (B11). The MMRS record's source (airisk.mit.edu, read 2026-10-28 'for scope, version')." }, { "id": "B2", "origin": "card", "role": "Overview card 2 at T1 (mit-ai-risk-repository-aio-20261029)", "card": "MIT AI Risk Initiative — AI Risk Repository Report updated (April 2025)", "url": "https://airisk.mit.edu/blog/new-version-of-the-ai-risk-repository-preprint-now-available", "kind": "Overview card (blog post, 23 April 2025)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 2 } ], "files": { "B2.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B2.html", "sha256": "437d6f93ad588640daf7dcd441149c333231918f2b6006afc6bef2c2ba31b822" }, "B2.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B2.txt", "sha256": "c6e93a76c8af929f15d9b19a36f9f7af53038591f08143ecc1bbde50df955b96" } }, "fetched": { "at": "2026-10-29T05:11:31Z", "http": 200, "url": "https://airisk.mit.edu/blog/new-version-of-the-ai-risk-repository-preprint-now-available", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B2.html", "raw_sha256": "437d6f93ad588640daf7dcd441149c333231918f2b6006afc6bef2c2ba31b822", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B2.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "AI Risk Repository Report updated (April 2025)", "card_snippet": "AI Risk Repository Report updated (April 2025) * What's new in this update Integration of 22 new AI risk frameworks, bringing the total number of included docum...", "unreachable": false, "names_operator": "yes (the Initiative in the page's navigation and featured posts)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "located by the card's title among the blog links on the home page; a different post of the same month, 'Repository Update: April 2025' (1 April 2025, Version 3), was first fetched in error and replaced by this one, whose title and opening match the card" }, { "id": "B3", "origin": "card", "role": "Overview card 3 at T1 (mit-ai-risk-repository-aio-20261029)", "card": "MIT Initiative on the Digital Economy — MIT Researchers Create an AI Risk Repository - MIT Initiative on the Digital Economy", "url": "https://ide.mit.edu/insights/mit-researchers-create-an-open-ai-risk-repository/", "kind": "Overview card (MIT IDE article with interview, 21 August 2024)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 3 } ], "files": { "B3.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B3.html", "sha256": "75803d35fbe47a7cc375583770716c6215115540dab56461725e4a8fbfd73f3e" }, "B3.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B3.txt", "sha256": "e3c9288651eb1a00de0155f607239a9b7ccad7c9d686246f81bd385d6b98fa44" } }, "fetched": { "at": "2026-10-29T05:09:25Z", "http": 200, "url": "https://ide.mit.edu/insights/mit-researchers-create-an-open-ai-risk-repository/", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B3.html", "raw_sha256": "75803d35fbe47a7cc375583770716c6215115540dab56461725e4a8fbfd73f3e", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B3.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "MIT Researchers Create an AI Risk Repository - MIT Initiative on the Digital Economy", "card_snippet": "Overview: MIT FutureTech’s AI Risk Repository catalogs 777 risks across 43 AI frameworks, providing source evidence, citations, and dual search taxonomies (doma...", "unreachable": false, "names_operator": "no (MIT FutureTech and the IDE named; the Initiative not named)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "the card's snippet opens 'Overview:' and is not a span of the page: it is a summary of the article in the surface's own words (the page says '43 AI frameworks … 777 risks'); found by WebSearch on title and site" }, { "id": "B4", "origin": "card", "role": "Overview card 4 at T1 (mit-ai-risk-repository-aio-20261029)", "card": "MIT FutureTech — FutureTech", "url": "https://futuretech.mit.edu/publication/the-ai-risk-repository", "kind": "Overview card (FutureTech publication page, v1 abstract, 14 August 2024)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 4 } ], "files": { "B4.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B4.html", "sha256": "d8b51fa0df9e925a06854ca527f75a3af622824b5b4dbe69c78dbbc5cbb58382" }, "B4.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B4.txt", "sha256": "56497bd8702d8e734b9c0ec62d1490916b86eea368652723019beb021622cfc4" } }, "fetched": { "at": "2026-10-29T05:09:25Z", "http": 200, "url": "https://futuretech.mit.edu/publication/the-ai-risk-repository", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B4.html", "raw_sha256": "d8b51fa0df9e925a06854ca527f75a3af622824b5b4dbe69c78dbbc5cbb58382", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B4.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "FutureTech", "card_snippet": "We construct our Repository with a systematic review of taxonomies and other structured classifications of AI risk followed by an expert consultation. We develo...", "unreachable": false, "names_operator": "no (MIT FutureTech is the site; the Initiative not named in the abstract)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "the card's snippet is a verbatim span of the v1 abstract on this page; a second FutureTech page for the published title exists (futuretech.mit.edu/publication/the-ai-risk-repository-a-meta-review-database-and-taxonomy-of-risks-from-artificial-intelligence) and was not the card" }, { "id": "B5", "origin": "card", "role": "Overview card 5 at T1 (mit-ai-risk-repository-aio-20261029)", "card": "MIT AI Risk Navigator — MIT AI Risk Navigator", "url": "https://www.airi-navigator.com/", "kind": "Overview card (the Navigator's home page; a sibling entity, a census candidate)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 5 } ], "files": { "B5.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B5.html", "sha256": "aff86294148022289c52942381f440afd6e3ae11bdfe2b3bad683cbc99b3576a" }, "B5.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B5.txt", "sha256": "0b1d110533f1a4e8d7b247f428287d614b4ab14db18635b86314f7fc96e8e6d3" } }, "fetched": { "at": "2026-10-29T05:09:25Z", "http": 200, "url": "https://www.airi-navigator.com/", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B5.html", "raw_sha256": "aff86294148022289c52942381f440afd6e3ae11bdfe2b3bad683cbc99b3576a", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B5.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "MIT AI Risk Navigator", "card_snippet": "MIT AI Risk Navigator Home Browse MIT AI Risk Navigator Explore the full landscape of AI risk in one place. The Navigator connects MIT's AI Risk Repository data...", "unreachable": false, "names_operator": "yes ('A project of the AI Risk Initiative at MIT FutureTech')", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "located through the link on airisk.mit.edu/blog/introducing-the-ai-risk-navigator and llms.txt; airisk.mit.edu/navigator is a separate stub page ('MIT Navigator') whose meta description reads 'MIT's AI Risk Initiative datasets', not the card's 'AI Risk Repository data…'" }, { "id": "B6", "origin": "card", "role": "Overview card 6 at T1 (mit-ai-risk-repository-aio-20261029); also an own page (the MMRS record's source / named by the task)", "card": "arXiv — [2408.12622] The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence", "url": "https://arxiv.org/abs/2408.12622", "kind": "Overview card (arXiv abstract page, v3); also an own page the task names", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 6 } ], "files": { "B6.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B6.html", "sha256": "79eb6de30ec0f0271dd82666c6d39e8d665aa0ae80770f2829c145f8aa0a7ecd" }, "B6.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B6.txt", "sha256": "7726a68799cf5a268bec448d6d294c2e9b7ac659edb01710675c824709192d30" } }, "fetched": { "at": "2026-10-29T05:08:15Z", "http": 200, "url": "https://arxiv.org/abs/2408.12622", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B6.html", "raw_sha256": "79eb6de30ec0f0271dd82666c6d39e8d665aa0ae80770f2829c145f8aa0a7ecd", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B6.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "[2408.12622] The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence", "card_snippet": "Overview: The AI Risk Repository catalogs and standardizes fragmented artificial intelligence risk terminology by analyzing 74 major risk frameworks containing ...", "unreachable": false, "names_operator": "no (the authors and MIT not named on the abstract page; the Initiative not named)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "the card's snippet opens 'Overview:' and is not a span of the abstract (the abstract says 'analyzed every major AI risk framework published to date-74 frameworks containing 1,725 distinct risks'; 'standardizes' is the snippet's word)" }, { "id": "B7", "origin": "card", "role": "Overview card 7 at T1 (mit-ai-risk-repository-aio-20261029)", "card": "MIT AI Risk Initiative — Repository Update: December 2025", "url": "https://airisk.mit.edu/blog/repository-update-december-2025", "kind": "Overview card (blog post, 4 December 2025)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 7 } ], "files": { "B7.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B7.html", "sha256": "8afe267cf8dd329a8f815610775a96692ebb9d45ac59f3d9c65da1fca0f687b3" }, "B7.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B7.txt", "sha256": "21849cbdbd3f025991ad652e10cafa5f1043b03956b38503c8360c6c9ec8ef08" } }, "fetched": { "at": "2026-10-29T05:08:15Z", "http": 200, "url": "https://airisk.mit.edu/blog/repository-update-december-2025", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B7.html", "raw_sha256": "8afe267cf8dd329a8f815610775a96692ebb9d45ac59f3d9c65da1fca0f687b3", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B7.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "Repository Update: December 2025", "card_snippet": "Foundational Challenges in Assuring Alignment and Safety of Large Language Models A Survey on Responsible LLMs: Inherent Risk, Malicious Use, and Mitigation Str...", "unreachable": false, "names_operator": "yes", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "the card's snippet is two of the post's listed frameworks' titles" }, { "id": "B8", "origin": "card", "role": "Overview card 8 at T1 (mit-ai-risk-repository-aio-20261029); also an own page (the MMRS record's source / named by the task)", "card": "MIT AI Risk Initiative — About", "url": "https://airisk.mit.edu/about", "kind": "Overview card (own page: the Initiative's About page)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 8 } ], "files": { "B8.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B8.html", "sha256": "b40b72f7dbd5c16c3d50da84dc12cbfc9974e71e5ffd84e3dc8f0a95305e3396" }, "B8.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B8.txt", "sha256": "a7b938e0aa7923a9b38c7dc508258053d9348c46e582916b770bbba25a0c39a1" } }, "fetched": { "at": "2026-10-29T05:08:14Z", "http": 200, "url": "https://airisk.mit.edu/about", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B8.html", "raw_sha256": "b40b72f7dbd5c16c3d50da84dc12cbfc9974e71e5ffd84e3dc8f0a95305e3396", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B8.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "About", "card_snippet": "No single dataset captures all of this, so we maintain several connected workstreams, each answering a different question. * AI Risk Repository What are the ris...", "unreachable": false, "names_operator": "yes", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "the task names this page as an own page; carried once, as a card" }, { "id": "B9", "origin": "card", "role": "Overview card 9 at T1 (mit-ai-risk-repository-aio-20261029)", "card": "LinkedIn — MIT AI Risk Initiative | LinkedIn", "url": "https://www.linkedin.com/showcase/mit-ai-risk-initiative/", "kind": "Overview card (LinkedIn showcase page, signed-out view)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 9 } ], "files": { "B9.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B9.html", "sha256": "d098a0edbdf235f7dbf902a32d0120638b6bbbcee91cca8c49da343bf60a1fe9" }, "B9.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B9.txt", "sha256": "ddf88ccb5ab1762adeb03456844e28541c86dae49db70ea72cd0f75bba7e71c6" } }, "fetched": { "at": "2026-10-29T05:09:26Z", "http": 200, "url": "https://www.linkedin.com/showcase/mit-ai-risk-initiative/", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B9.html", "raw_sha256": "d098a0edbdf235f7dbf902a32d0120638b6bbbcee91cca8c49da343bf60a1fe9", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B9.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "MIT AI Risk Initiative | LinkedIn", "card_snippet": "The initiative includes the MIT AI Risk Repository, a living database of more than 1,700 AI risks, the AI Incident Tracker, a collaboration with the Responsible...", "unreachable": false, "names_operator": "yes", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "the card's snippet is a span of a post on the page (the post's closing paragraph, repeated on two posts); linkedin.com/company/mit-ai-risk-initiative and /showcase/mit-ai-risk-repository/ both redirect here" }, { "id": "B10", "origin": "card", "role": "Overview card 10 at T1 (mit-ai-risk-repository-aio-20261029)", "card": "ScienceDirect.com — The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence - ScienceDirect", "url": "https://www.sciencedirect.com/science/article/pii/S2666389926000267", "kind": "Overview card (the published article, Patterns, doi:10.1016/j.patter.2026.101517)", "surfaced_by_T": [ { "slug": "mit-ai-risk-repository-aio-20261029", "card": 10 } ], "files": { "B10.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B10.html", "sha256": "c307c5915d333b3f22fa9a6b3f97f94aae98142cdb904ccb5b6560b40adf3666" }, "B10.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B10.txt", "sha256": "2b435a31e901954b79bb7f37769288b62e8491182b32066550110e229fd63a6b" } }, "fetched": { "at": "2026-10-29T05:09:55Z", "http": 403, "url": "https://www.sciencedirect.com/science/article/pii/S2666389926000267", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B10.html", "raw_sha256": "c307c5915d333b3f22fa9a6b3f97f94aae98142cdb904ccb5b6560b40adf3666", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B10.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'; text by stripping scripts, styles and tags (the workspace's own extractor, scratchpad h2t.py), whitespace kept by line" }, "card_title": "The AI risk repository: A meta-review, database, and taxonomy of risks from artificial intelligence - ScienceDirect", "card_snippet": "* Practical implications for technology management. Our AI Risk Repository provides concrete tools for managing AI's societal impact across multiple stakeholder...", "unreachable": true, "names_operator": "— (unreachable)", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)", "note": "unreachable: located through the DOI on the arXiv page (B6: 'Patterns 101517 (2026)', doi:10.1016/j.patter.2026.101517 → linkinghub pii S2666389926000267); www.sciencedirect.com answered HTTP 403 twice (05:09 and 05:09:55 UTC; B10.html is the 403 body), and the publisher's other host, cell.com/patterns, answered 403 (a challenge page) once; no further retry. The snippet stands as the card's only text (B10.txt holds the card's title and snippet). Whether the snippet's 'Practical implications for technology management' heading is in the Patterns article could not be checked" }, { "id": "B11", "origin": "own_page", "role": "own page, not surfaced by T; admitted beside the cards by the operator's ruling of 2026-10-29", "card": "own page: the repository's own page (AI Risk Repository, airisk.mit.edu/risks)", "url": "https://airisk.mit.edu/risks", "kind": "own page: the repository's own page (AI Risk Repository, airisk.mit.edu/risks)", "surfaced_by_T": "no", "files": { "B11.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B11.html", "sha256": "a5bd76e2fcecfd1035d1812eb1a9df6d30ecabf8f42120de43ce6230b7fd5442" }, "B11.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B11.txt", "sha256": "2b705e1a289760feea4a0f2dd66d3c742e4c3a5fa48079fa60eaf9f266cf3228" } }, "fetched": { "at": "2026-10-29T05:11:05Z", "http": 200, "url": "https://airisk.mit.edu/risks", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B11.html", "raw_sha256": "a5bd76e2fcecfd1035d1812eb1a9df6d30ecabf8f42120de43ce6230b7fd5442", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B11.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'" }, "names_operator": "yes ('MIT AI Risk Initiative', 'MIT FutureTech')", "names_archive": "no", "note": "not cited by the MMRS record, whose URL airisk.mit.edu/ now opens on the Initiative's home (B1); admitted as the repository's own page because B1 links it as 'All Risks' and llms.txt (B12) names it the repository's canonical page; not surfaced by T" }, { "id": "B12", "origin": "own_page", "role": "own page, not surfaced by T; admitted beside the cards by the operator's ruling of 2026-10-29", "card": "own page: airisk.mit.edu/llms.txt, the Initiative's file for language models and AI agents", "url": "https://airisk.mit.edu/llms.txt", "kind": "own page: airisk.mit.edu/llms.txt, the Initiative's file for language models and AI agents", "surfaced_by_T": "no", "files": { "B12.html": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B12.html", "sha256": "039e65007ed8770062e34d9730acc7484764b755b80165a3e0de9561489cf0c6" }, "B12.txt": { "path": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B12.txt", "sha256": "039e65007ed8770062e34d9730acc7484764b755b80165a3e0de9561489cf0c6" } }, "fetched": { "at": "2026-10-29T05:11:06Z", "http": 200, "url": "https://airisk.mit.edu/llms.txt", "raw": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B12.html", "raw_sha256": "039e65007ed8770062e34d9730acc7484764b755b80165a3e0de9561489cf0c6", "text": [ "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/B12.txt" ], "how": "curl -sSL -m 30 -A 'Mozilla/5.0'" }, "names_operator": "yes", "names_archive": "no", "note": "not cited by the MMRS record; admitted as the entity's own page, the site's statement addressed to the composition layer itself; a source carrying its own summary policy (§3.5): the policy audits the extraction and does not substitute for it, and it is extracted by the same rule as every other source; not surfaced by T; plain text, so .html and .txt are the same bytes" } ], "wikipedia_record": { "id": "W1", "origin": "wikipedia", "result": "no article", "checked": "2026-10-29 05:10–05:11 UTC", "requests": [ { "url": "https://en.wikipedia.org/w/index.php?title=MIT_AI_Risk_Repository&action=raw", "http": 404, "file": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/W1-raw.html", "sha256": "cd877cb2f795da2681718566942344434b7b71c4ed877be187cbccc126ed5f60" }, { "url": "https://en.wikipedia.org/w/index.php?title=AI_Risk_Repository&action=raw", "http": 404 }, { "url": "https://en.wikipedia.org/w/index.php?title=AI_risk_repository&action=raw", "http": 404 }, { "url": "https://en.wikipedia.org/w/index.php?title=MIT_AI_Risk_Initiative&action=raw", "http": 404 }, { "url": "https://en.wikipedia.org/w/index.php?title=MIT_FutureTech&action=raw", "http": 404 }, { "url": "https://en.wikipedia.org/wiki/MIT_AI_Risk_Repository", "http": 404, "file": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/W1.html", "sha256": "88df00119de6e790135f1ab93c8427d9950ac5c09133df6a0d4adb60b7c216a5", "says": "Wikipedia does not have an article with this exact name." }, { "url": "https://en.wikipedia.org/w/api.php?action=query&list=search&srsearch=%22AI%20Risk%20Repository%22", "http": 429, "note": "rate-limited; replaced by Special:Search" }, { "url": "https://en.wikipedia.org/w/index.php?search=%22AI+Risk+Repository%22&title=Special%3ASearch&profile=advanced&fulltext=1&ns0=1", "http": 200, "file": "datasets/negative-of-the-negative/v2/intake/mit-ai-risk-repository-20261029/W1-search.html", "sha256": "c9b3bdabd488fe720cc6bbb07cd760ff5a4bd7656d284532cf409c3a64c92925", "says": "The pages \"AI Risk Repository\" and '\"AI Risk Repository\"' do not exist; one result, the article 'Voiceverse NFT plagiarism scandal' (last edited 10 October 2026), whose text reads 'The MIT AI Risk Repository catalogued the incident as an example of \"economic and cultural devaluation…'" } ], "redirect": "none (no title variant resolves; no redirect page)", "note": "No article for the entity, its operator (MIT AI Risk Initiative) or its group (MIT FutureTech) exists on English Wikipedia on 2026-10-29: the raw-wikitext requests for five titles answer HTTP 404, the rendered page answers 404 with 'Wikipedia does not have an article with this exact name.', and a full-text search for the quoted name returns one article that cites the repository's classification of one incident (Voiceverse NFT plagiarism scandal), not an article on it. The fact is a property of the row: the field has no Wikipedia origin; the one mention is recorded here and not ledgered (it is not the entity's article and was not surfaced by T)." } } }